Posts in Business (20 found)
Stratechery Yesterday

IBM Misses, IBM’s Mainframe Moat, IBM’s Many AI Problems

IBM announced preliminary results that spooked the software market generally; this is a story, however, specifically about IBM and its mainframe franchise.

programming Programming hardware Hardware business Business
0 views
Michael Lynch Yesterday

Refactoring English: Month 19

Hi, I’m Michael. I’m a software developer and founder of small, indie tech businesses. I’m currently working on a book called Refactoring English: Effective Writing for Software Developers . Every month, I publish a retrospective like this one to share how things are going with my book and my professional life overall. At the start of each month, I declare what I’d like to accomplish. Here’s how I did against those goals: I improved the website a bit, but it could use more polish. I adapted my chapter on design docs to a free excerpt . It did well on Lobsters and Reddit , but it flopped on Hacker News. I was surprised at how positive the reaction was to the design docs chapter. Generally, when I talk to developers about design docs, their main reaction is that they hate design docs and everything about them. The comments on my post were refreshingly supportive of design docs in general and my recommendations in particular. I got stuck for a while on the great AI blockade , but I pushed through by thinking more critically about splitting up large features and being less precious about code quality. In this case, done is better than perfect. June was the best month of book revenue since the initial crowdfunding launch. The increase in visitors was because of my excerpt about design docs . For the last few months, the Refactoring English website has listed my book as almost complete in early access. I was curious to see what the sales impact would be of going from an almost complete book to a fully complete book, so I looked at weekly sales: Marking the book as complete didn’t have an obvious impact on weekly sales, but what if I look at the daily averages? Okay, so there was a slight increase after I marked the book as complete. I was also curious whether Americans, in particular, bought at higher rates after I finished the book. I get email notifications every time someone purchases the book, and it seemed like more of my sales were from customers paying the US price, but I hadn’t measured carefully. I checked the data to see if that was true: Interesting! Completing the book had no impact on sales for customers purchasing with regional pricing, but customers purchasing in USD purchased at a 20% higher rate in the three weeks after the book was complete. I didn’t include sales after I published my latest excerpt because that obviously changes the numbers a lot, so let me treat that as its own category: But that’s always a little skewed because Americans make up the largest share of my readers. What if I normalize revenue per visitor? Oh, that’s a switcheroo. By normalizing per visitor, it flips the story. Now, it’s the Americans that buy at the same rate for a finished vs. unfinished book. The readers outside the US are the ones spending about 20% more per visitor on the completed book. I’m not sure how to use this information, but it did satisfy my curiosity. I’ve asked readers for feedback about my book in the past, and some readers gave enthusiastic feedback, but they were a small minority. I thought it would be fun and helpful to make a web-based feedback app that allows readers to leave notes as they read the book. It seemed like something I could knock out in a week or two. And now, two short… months later, I’ve got it up and running! A demo of my book feedback tool, where readers can leave me feedback directly in the book, and I can reply. My feedback tool has only been live for a few days, but it does seem to encourage readers to give more feedback. One reader just finished the book and cited the feedback app as one of his favorite parts of the experience, so that was neat. About once a year, I ask myself: where does all my time go? This question comes up for me whenever I’m focused on a project, but it’s not progressing as quickly as I expect. Here’s me asking myself this question a few times over the years: This time, I thought, “Maybe I should use a time tracking tool.” About 15 years ago, I tried a time tracking tool called RescueTime. I didn’t find it that useful, but I thought maybe I’d keep at it for a few weeks and see what happened. Then, I realized I was letting a random company collect data about every window that appeared on my screen, and I promptly uninstalled RescueTime. I was wishing for an open-source version of RescueTime, when I thought, “Wait, there probably is one.” And there is. It’s called ActivityWatch . It’s open-source and privacy-first. It records all your window and browsing activity, but the data all stays local to your machine. The problem is that ActivityWatch is way less polished than RescueTime. I couldn’t understand at all what the timeline was trying to show me: I couldn’t understand the timeline in the official ActivityWatch web interface. You’re supposed to assign rules to tell ActivityWatch how to categorize your activities, but I found that UI difficult to use as well: I found the categorization in the official ActivityWatch web UI difficult to use. I was about to give up on ActivityWatch, and then I thought, “Well, the data collection part probably works. What if I vibecode my own frontend?” So, I did , and it was pretty easy. I’m starting with a command-line tool, but I plan to expand it to a web app. To use my custom ActivityWatch frontend, I create a config file to categorize activities based on app name, window title, and/or URL: And then the output looks like this: So far, the data is interesting, but the biggest challenge is that it’s hard to categorize all of my activities automatically. For example, I can add a category for browsing Wikipedia, but am I doing it as part of legitimate work on my book? Or did I just go down a rabbit hole, and I’m suddenly reading about inventors killed by their own inventions ? Refactoring English had its second-best month of sales. I examine my sales numbers to see whether people are more likely to purchase a complete book as opposed to an almost-complete draft. I completed my book feedback tool. I’m trying a new tool to track my time. Result : Spent about three hours improving the website Result : Got 17.5k unique readers. Result : The tool is up and running. Finished the Refactoring English feedback tool. Made fixes to the Refactoring English ebook for consistency and EPUB compatibility. Made a demo video for Little Moments . I’m quite proud of the silly photos in this. Customers don’t care as much as I’d expect about the difference between a 100% complete book and an almost-complete book. Readers do purchase the finished book at higher rates, but the effect is pretty small when you control for number of website visitors. Pitch to 5 podcasts to talk about Refactoring English . Attract 30k unique readers to the Refactoring English website. Wrap up early access, and declare the 1.0 release of my book.

books Books writing Writing business Business
0 views
DHH Yesterday

Dell is on a roll with the XPS

We've been buying servers from Dell since the 2000s at 37signals, but I was never too impressed with their personal computers. They either felt cheap or enterprisey to me. Like they were made exclusively for people who are handed standard-issue laptops by corporate, and not something discerning techies would buy with their own money. But the new XPS line has completely changed my perception. I've now spent several months with the 2026 XPS 14 and 16, and last week I added the MacBook Neo-fighting XPS 13, and all I can say is that these machines are fantastic! Great chips, great screens, great build quality. Superb packages. Which is very satisfying to see because there are few American business leaders I respect more than Michael Dell. He's been running his company for over forty years now, and he's still calling the shots! So to see the company pull a turnaround like this, so many years into its run, is very inspiring. I've written about the XPS 14 before, and as I noted back in April, a good portion of the credit for these new Dell machines being really good belongs to Intel. The 18A process is paying big dividends for both companies (and the rest of the PC makers). But Dell could still have stuck these chips into forgettable machines, and I wouldn't have had any interest. In fact, they did! Just last year, for the 2025 model year, they shipped new XPS machines with awful capacitive-touch function and esc keys. Two years after Apple had finally thrown in the towel on the ill-fated Touch Bar on their MacBooks! Dell also killed the XPS branding last year, and went with the truly uninspired Plus/Premium/Pro copycat branding. Like some cheap Chinese knockoff. It was embarrassing, to be honest. But unlike Apple, which introduced that cursed Touch Bar back in 2016, and then crammed it down everyone's throat for seven long years, Dell rebooted this nonsense almost immediately. Gave us back real function and esc keys, and revived the XPS branding. You could argue that they should have learned from Apple's mistakes to avoid their own, but the next best thing is surely a quick reversal. And what a reversal it's been. As I said, I've spent months using an XPS 14 as my main machine. It's been so good I even gave up on using a dedicated desktop machine. Now I just run everything off the XPS 14, connected to an Apple XDR 6K 32" (nobody has yet managed to beat this, and I've owned it for years). It's a great, simple setup. The XPS 14 is an expensive machine, though. Not more so than its direct competitors, but still, at $2,799 for the 358H/32GB/1TB/OLED unit, it's a lot. I'd spend that in a heartbeat, but not everyone is going to drop that kind of cash on a laptop. Especially if they already have a powerful desktop. That's where the new XPS 13 comes in. It's part of the PC industry's answer to Apple's new MacBook Neo, which analysts all thought would catch the other side flat-footed. Well, surprise, it didn't! Apple charges $699 for an 8GB RAM/256GB SSD Neo, whereas Dell wants $699 for 8GB RAM/512GB SSD, and even offers a 16GB RAM/512GB SSD version for $899 (there's no RAM upgrade possible for the Neo). But matching Apple on specs and price wasn't the surprise; it was besting them with a nicer screen and keyboard, and meeting them on build quality. The XPS 13 has a great 120Hz screen (something you don't even get on a MacBook Air at twice the money!), a superb keyboard w/ backlighting (also missing on the Neo!), and weighs 20% less at just 1 kg with every bit as nice an aluminum chassis. Now I'd forgive anyone their skepticism about 8GB RAM and Windows. Microsoft isn't exactly known for creating a responsive operating system on modest specs these days, but who cares, we have Linux! Of course, I've been running Omarchy on this thing for the past week, and it's frankly fantastic. As long as you understand the limitations! The Intel Wildcat CPU uses the same performance cores as the full Panther Lake chip, so single-threaded snappiness is all there, but it only has two of those, and then another four low-powered cores. So six total, but not a mix that's conducive to running big multi-core workloads, like local CI. This is where the XPS 13 meets the moment. As the agent craze has been taking over software development, you might have seen any of the many memes about half-cracked laptops, just so the agents won't halt with a closed lid. The obvious answer is of course to run these agents off a home server in the closet, connect them to something as slim and light as an XPS 13 over Tailscale, and then control it all over SSH. Used like this, you get a machine that runs a browser as fast as anything on the PC (thanks to those full-speed performance cores) while costing a fraction of a new top-spec machine, and having better close-the-lid ergonomics. Win-win-hurray. When I posted my enthusiasm on X about this new XPS 13, I got at least three replies with "Is this an ad???". No. This is not an ad. I bought the XPS 13 with my own money, and frankly, you couldn't pay me any sum to use a laptop I didn't like. I did try Dell's laptops a few years back, didn't like what I saw, and ended up spending a few years using Framework computers instead (they're still great too). I'm simply excited that the PC isn't giving up without a fight. That Linux has been on a run among early adopters. That companies like Intel and Dell are here to keep Apple honest. Competition is great. It was Apple's M chips that rejuvenated the laptop market, and they held a supreme lead for years. So it's lovely to see Intel, Dell, and others actually being ready to meet the challenge from the low-cost Neo right out of the gate. So I tip my hat, once again, to Michael Dell. Forty-plus years at the helm, not too proud to pivot quickly, and now the maker of my favorite Linux laptops. Well done, sir.

business Business Electronics hardware Hardware
0 views
Stratechery 2 days ago

The OpenAI Super App, ChatGPT = Codex, Whither Chat

OpenAI has refashioned Codex as the new ChatGPT; is the company abandoning the chat category they pioneered?

ai AI business Business
0 views
マリウス 2 days ago

The Day WhatsApp Goes Dark

Note: As usual, tl;dr at the end. Tomorrow morning, WhatsApp goes dark, and it’s not just a short downtime, but it is a termination of the service. The servers turn off, the domains don’t resolve anymore and no mobile client is able to connect. Have you ever asked yourself what would happen in that case? What if WhatsApp actually went dark? Obviously, nobody really knows what would happen in such a case, because we haven’t experienced that situation (yet), but even though the closest analogues like the six-hour Meta outage in October 2021, and Brazil’s 12-hour court-ordered shutdown in December 2015 were measured in hours, not days, those already produced effects that journalists described as “apocalyptic” . We can try to extrapolate what happened throughout events like those to see what “global catastrophe scenario” could theoretically look like. Because whether you believe it or not, WhatsApp is more than just a messenger , and one example that makes this pretty obvious came from the Forbes editor José Caparroso , who wrote during the 2021 blackout that … Latin America lives on WhatsApp . I am surprised by so many people underestimating how catastrophic this downfall has been. But before we dive into this thought experiment, however, it’s worth establishing what we’re actually talking about, as readers in most of Europe and North America underestimate WhatsApp by an order of magnitude, primarily because in those markets it functions as one platforms among many. That is, however, not how the rest of the planet works. Note: This thought experiment is not only based on some abstract numbers and studies, but upon my own experience of how WhatsApp is being used in e.g. the global south on a day to day basis. During my travels I think I’ve pretty much “seen it all” , with for example broadband technicians taking photos of the stickers on the backside of WiFi routers/modems, that show the hardware address and login credentials (on their phones), and sending them via WhatsApp to themselves, only so they can open them on WhatsApp Web (on their work laptops), in order to upload them into the ISP’s technical service portal. It is frankly mind-boggling what sort of tasks WhatsApp has become a Swiss army knife for in those countries, whether it’s as a file transfer platform for sensitive documents, or as a full-blown hotline for critical services and infrastructure. Let’s start by understanding the sheer scale of WhatsApp . The Meta owned and operated messenger has roughly 3.3 billion monthly active users as of early 2026, which is about 40% of every human alive, and somewhere north of 60% of every human with a smartphone. The platform processes more than 100 billion messages per day , out of which around 7 billion are voice messages. On top of that, users place around 5.5 billion voice calls and 2.4 billion video calls per month , which boils down to more than 2 billion minutes of voice and video traffic every 24 hours. To put this in perspective, the global SMS network, at its peak in 2012, handled about 23 billion messages per day across every carrier on Earth. WhatsApp does four to five times that volume on its own, every day, on a service that is (at least at the consumer layer) “free” . However, if we look deeper into the country-level breakdown, it becomes clear that WhatsApp usage isn’t evenly distributed across the globe. India has between 535 million and 596 million monthly active users , and regardless of whether we pick the higher number or we stick with the more conservative estimate, it is the largest single national user base on any messaging platform anywhere. Brazil has about 148 million users, and the app is installed on roughly 99% of the country’s smartphones. And 93% of those users open the app daily . Indonesia has about 112 million users, with WhatsApp being the leading messaging platform in the country, and in Zimbabwe WhatsApp alone accounts for roughly 44–50% of all mobile internet traffic . In Lebanon more than four in five adults use it , making it the dominant communications channel during multiple national crises. In a great many countries, WhatsApp is not simply a service on the internet, it actually is the internet for most practical purposes. WhatsApp Business now has more than 200 million businesses on the platform globally , with around 50 million small and medium-sized enterprises using it as their primary customer channel. In India and Brazil, roughly 80% of small businesses use WhatsApp to communicate with customers. In Brazil specifically, 96% of businesses rate WhatsApp as their primary communication tool, and a joint study by Fundação Getulio Vargas and Sebrae , Brazil’s main small-business support organisation, found that 70% of Brazilian small companies rely on the Meta -owned trinity ( WhatsApp , Instagram , Messenger ) as their marketplace. Globally, around $45 billion in commerce is expected to flow through WhatsApp in 2026 . Click-to-WhatsApp advertisements alone generate roughly $10 billion per year for Meta . About 175 million customers send messages to WhatsApp Business accounts every single day. And then there’s payments. In India, WhatsApp Pay is a small player in the UPI with about 67 million transactions per month against UPI’s 18 billion monthly volume, but in absolute terms, that’s still an enormous number of transactions. In Brazil, WhatsApp Pay is integrated with local card and bank rails and is used by transit operators ( Vai de Bus , for instance, sells passes via WhatsApp ), banks, and merchants. Across Africa, fintech overlays on WhatsApp , like Finnova in Nigeria, or Azza in Nigeria, Kenya, and South Africa, are processing crypto and conventional payments at significant volumes. Besides being a chat platform, a marketplace and a payment processor, WhatsApp is also being used as critical clinical infrastructure across the global south. A three-year programme at UCLA’s David Geffen School of Medicine paired subspecialists in Los Angeles with clinicians at Partners in Hope Medical Center in Lilongwe, Malawi, via WhatsApp groups. 89% of submitting clinicians and 71% of expert respondents reported that the case discussions improved medical education and patient outcomes. In the Eastern Cape of South Africa , WhatsApp groups serve as the primary continuing-medical-education channel for HIV and TB management in rural clinics where specialists are days away. In Haiti, WhatsApp groups coordinate emergency department operations at Hôpital Universitaire de Mirebalais , including mass-casualty alerts, security updates, and clinical decisions. In Zambia, IntraHealth International runs nurse and midwife mentoring networks over WhatsApp . In Brazil, the link between Zika virus infection and microcephaly was tracked partly through WhatsApp groups of paediatricians comparing cases. Another critical field that runs on Meta ’s infrastructure is disaster response. The World Bank documented that during 2014’s Cyclone Hudhud in Andhra Pradesh, India , the Public Works Department restored connectivity to a 1.8-million-person city primarily by coordinating engineers through a closed WhatsApp group with the District Magistrate in it, without any formal meetings and orders, which ultimately led to most roads becoming functional within three to four days. During the 2023 Turkey earthquakes, volunteer-formed WhatsApp networks processed 5,800+ messages in one week for needs assessment and rescue, and in Syria, the White Helmets have run an emergency dispatch system over WhatsApp since 2021, because the country’s emergency number infrastructure is largely destroyed and WhatsApp ’s compression algorithms work where almost nothing else does. It’s not just individual organisations, but even whole governments are dependent on Meta . Buenos Aires for example ran a COVID-symptom triage chatbot on WhatsApp , and Lebanon’s public health ministry launched an automated WhatsApp service in April 2020 to disseminate updates on the pandemic. India, on the other hand, offers metro tickets, government services, and bill payments through WhatsApp chat interfaces . On top of that, for example, the Philippines’ UAE consulate operates consular emergency hotlines on, you guessed it, WhatsApp . Last but not least, there’s migration. Roughly a quarter-billion people live outside their country of birth. Most of them use WhatsApp as their primary connection to family, because international SMS is expensive and unreliable and Skype is, well, dead. Multiple peer-reviewed studies on Trinidadian , Pakistani, Ghanaian , Polish, and Kenyan diasporas also converge on the same finding of WhatsApp being the primary technology of transnational family life in 2026. So to go back to our initial thought, let’s imagine WhatsApp shutting down in an instant, with this dependency graph in mind. What follows is a hypothetical scenario sketched from the documented impacts of past (shorter) outages, scaled up by the duration and finality of the event, and informed by the dependency layers described above. It’s a scenario and not an actual prediction. The shutdown hits during European afternoon, which means American morning, Indian evening, East African afternoon, and Indonesian late evening. The first signals show up on Downdetector and on non- Meta competitors. In 2021, the six-hour outage generated 14 million reports inside the first few hours, but this time the number is likely much larger. Behaviour inside the first hour is uneven and largely confused. In most places, users assume it’s a routing problem, a local carrier issue, or a phone bug. They restart the app, then their phone, then their router, then they check Twitter X , Instagram , TikTok , Telegram , maybe Signal , or Facebook Messenger , depending on what they have installed. Telegram and Signal both see app-store download spikes within the first 30 minutes, as it happened during the 2021 outage, with Signal reportedly adding “millions” of users that day . The first noticeable failures show up in commerce. A food-truck operator in São Paulo who takes orders via WhatsApp can no longer receive them. A small clothing brand in Mumbai whose entire sales pipeline runs through Click-to-WhatsApp advertisements sees its ad spend continuing to bill while the conversation endpoint returns errors. In Hong Kong, a logistics coordinator who confirms container pickups via WhatsApp loses the day’s confirmation chain. In Idlib, Syria, the White Helmets dispatch room realises within minutes that emergency calls are not coming in, and civilians have no fallback channel. It is likely that three things start happening in parallel. First, mass migration to apps like Telegram , Signal , and to a lesser extent Messages ( iMessage ), Viber , and Line . Signal ’s servers, which are run on a fraction of WhatsApp ’s infrastructure, are not designed for an inrush of hundreds of millions of new accounts and start to degrade in some regions. Telegram , which has spent a decade preparing for exactly this scenario, holds up better but still struggles with its own issues. Ultimately none of the alternatives are suitable for the people who had built their workflows on WhatsApp . The second thing that happens is commercial collapse , which is the biggest 12-hour story, but still largely invisible from Western media. In Brazil, Indonesia, Nigeria, India, Pakistan, Bangladesh, Vietnam, Mexico, and probably 50 other countries, the small businesses that route everything from orders and prices, and photos of goods, to delivery confirmations, and payments, through WhatsApp have lost their primary revenue channel. A clothing brand in Ireland reportedly lost thousands of euros in a single afternoon during the 2021 six-hour outage. Multiply this by twelve hours and by the entire tail of informal commerce that lives on the platform and the figure runs into the billions. The third thing is health-system stress . Group consults that normally take an hour over WhatsApp become almost impossible. The Eastern Cape HIV-management network in South Africa, the Malawi-UCLA clinical link, the Haitian ED coordination groups, the Zambian rural-nurse mentoring channels, all degrade simultaneously, and while mortality consequences are not yet visible, they are happening nonetheless. In several countries, government officials begin issuing statements through whatever channel is still functioning. After the first 24 hours it becomes clear that the impact this situation has is roughly inversely proportional to a country’s investment in alternative digital infrastructure. The United States and Western Europe are mildly inconvenienced, and India is moderately disrupted, mainly because the country has built duplicate rails, hence UPI runs over many apps. After all, SMS still works, alternative payment apps exist, and government services have their own portals. However, countries like Brazil, Argentina, Mexico, and most of sub-Saharan Africa, on the other hand, are in serious trouble. In Brazil, by the end of day one, the financial press is comparing the situation to a partial shutdown of the national payments system. Pix transfers still work, as those run over the central bank’s infrastructure and not WhatsApp ’s, but the merchant-customer communication layer that drives Pix transactions for millions of small operators is offline. The same is true in Argentina, where the inflation-driven culture of constant price renegotiation between vendors and customers happens, in practice, almost entirely on WhatsApp . Another area that starts to fail is migrant remittance. People working in the Gulf, North America, or Europe typically coordinate transfers with their families via WhatsApp , where they confirm the recipient’s details, send screenshots of receipts, or sometimes route the money through informal Hawala -style networks where trust is established and maintained by daily messaging. These workflows don’t fail completely on day one, but they slow and break in ways that don’t show up in formal remittance statistics for another week or two. In Latin America, the first major political consequence appears in the form of misinformation that previously circulated within closed WhatsApp groups , which now has nowhere to go and starts spilling onto other platforms. By the end of day one, more than 100 million people have created Signal or Telegram accounts. Both apps experience their first significant performance degradation events. The labour-market consequences start showing up. In India, where WhatsApp is the de facto recruiting and onboarding tool for huge segments of the informal economy, gig workers can’t be reached for shifts. Delivery platforms like Swiggy , Zomato , Dunzo , and their international equivalents, see their dispatch coordination degrade. Some of these companies have parallel in-app messaging, but many have leaned hard on WhatsApp because it was cheaper. Schools also begin to feel it, because in many countries, including India, Brazil, South Africa, Kenya, Nigeria, the Philippines, Indonesia, and much of the Middle East, parent-teacher communication runs over WhatsApp groups. Two days in, schools that have not made the switch to other channels are operating partially blind, and parents are not getting closure notifications, transport updates, fee reminders, or exam schedule changes. In countries with weak alternative communication infrastructure, the second-order effect is mid-week absenteeism as parents simply don’t know whether school is open. On top of it all, Healthcare is also heavily impacted. For example, the Haiti emergency-department-style coordination groups have now had 48 hours to find alternatives, and they have, mostly, but the transition has costs. Case discussions that were asynchronous and 24/7 on WhatsApp are now synchronous and harder to schedule, and rural clinicians in places like the Eastern Cape, Lilongwe, or the highlands of Nepal are once again practising in the relative isolation that WhatsApp ’s group-call and group-message features had alleviated. In several documented studies, isolation correlates with diagnostic delays and worse patient outcomes. In Syria, the White Helmets switch to a patchwork of Signal , SMS where it works, and physical runners, and response times degrade significantly. At this point things start to get political. In a number of countries, including Brazil, India, Indonesia, Nigeria, the Philippines, and South Africa, the question stops being “what is Meta doing” and starts being “why did we let one foreign company become this central” . Telecom operators in several countries pitch the moment as an opportunity to push their own messaging products, most of which have been moribund since 2014, but the pitches fail because nobody trusts the carriers, because those carriers have been quietly delighted to see WhatsApp gone, given that it eroded their SMS and voice revenue for a decade. In a few markets, regulators float emergency-decree-style proposals to nationalise messaging infrastructure or build sovereign alternatives. And while most of these proposals are clearly performative, some are not. India and Brazil both have working national digital identity and payments stacks that could, in principle, host a public messaging layer. It remains to be seen, though, whether the political will to build one persists past the first month. Public health authorities in Lebanon, Buenos Aires, the Philippines, and several African countries are now running emergency communication operations across multiple fallback channels. None of them work as well as WhatsApp did and things like vaccination schedules are missed, and appointment reminders fail. Some clinics see patient no-show rates rise by 30–40% versus baseline. Not because WhatsApp is superior to its competitors, but simply because humans need a long time to adjust to the alternatives that are being put in place. Also, crime patterns shift in interesting ways. A Conflict Sensitivity Resource Facility report on South Sudan, and PeaceRep work on Somalia, both documented that WhatsApp groups were used for both peace-building and for coordinating violence. Removing the platform doesn’t remove either function, as both migrate to other channels, but the migration takes time, and during the transition, coordination of all kinds becomes harder. In several markets, online ad spend collapses because Click-to-WhatsApp ads (a $10B/year business) have no destination, and Meta ’s stock price has already done what you’d expect it to do. The migration to alternatives, mostly Telegram and Signal , with regional pockets going to Line , KakaoTalk , WeChat , Messages ( iMessage ), RCS , and a long tail of smaller apps, has now hit critical mass in most of the world. The migration has not been clean, and group chats with over 200 members have, in practice, often migrated as group chats with around 40 members, because not everyone moved at the same time or to the same app. For business communication, the new world is as fragmented as it gets. A Brazilian shopkeeper who used to take all orders on WhatsApp now has to manage Telegram , Signal , Instagram DMs (still up, but reduced after Meta ’s reputational damage), and SMS. Customer-acquisition costs rise, and customer-retention drops, and several reporters publish stories on small businesses that have permanently closed. For healthcare, the migration is more orderly because the user base is smaller and more motivated. Most major peer-support networks, like the Malawi-UCLA , the Eastern Cape HIV , the Zambia nursing , and the Haiti emergency have stable new homes. The five-day disruption produced measurable degradation, and it is not yet possible to quantify the mortality and morbidity impact. In Syria, the White Helmets have built a partial replacement on Signal and on a custom dispatching tool that their engineers had been prototyping. It works less well than what they had, because the compression behaviour that made WhatsApp viable in low-bandwidth, intermittently-connected environments is hard to replicate. Hence, some dispatches are now arriving via paper notes. Not because decentralized mesh networks don’t exist, but simply because nobody in these organizations has the expertise to implement these alternatives, especially within such a short period of time. The first credible economic estimates of the shutdown’s cost reach the tens of billions of dollars and continue to rise. The estimates are dominated by long-tail effects in emerging markets that are hard to measure precisely. A week in, the question has shifted from “When does WhatsApp come back?” to “What does the world look like without it?” and a growing fraction of the user base assumes it isn’t coming back, so behaviour begins adapting accordingly. Several governments, including Brazil, India, and the EU as a bloc, have announced formal investigations or task forces into how to prevent this from happening again. As usual, however, none of them will produce anything actionable within years. The longer-term effects, that you can already see the shape of by day seven are a measurable productivity hit in emerging markets, particularly for informal-sector businesses, a consumer trust impact across the entire Meta product family, a wave of WhatsApp-replacement startups, most of which will fail due to network effects and generally bad engineering, and the painful realisation that a free product is not the same thing as a public good. Some estimates from prior outage studies suggest that a six-hour WhatsApp outage cost the global economy hundreds of millions of dollars per hour in lost SME activity, weighted heavily toward Latin America, South Asia, and Africa. Extrapolated over seven days and weighted for cascading effects, the seven-day damage is in the tens of billions, possibly higher. This thought experiment is not about Meta eventually shutting down WhatsApp , as it almost certainly won’t do so on its own, given how big of a lever the platform is for the company. In fact, Meta is moving in the opposite direction, as it is building WhatsApp Business into a $45 billion commerce platform, integrating it with payments, and turning ads into one of its fastest-growing revenue lines. WhatsApp is too valuable to Meta to switch off voluntarily, and the regulatory regimes in the countries that depend on it most are nowhere near coordinated enough to force a switch away from it or even just ban it outright. The point is that we have built a planet-spanning piece of communication infrastructure whose ownership, governance, and continuity are concentrated in a single American corporation, that is led by people with questionable values and beliefs, which all in all is a state of affairs that has no historical precedent. Sure, there are other US-based companies that “own digital communications” , like Twitter X and many others, albeit I’d argue that none of those platforms are so engrained into everyday life across many (predominantly developing) nations as WhatsApp is today. The closest analogue in scale is the global SMS network of the early 2000s, which, however, was federated, run by hundreds of carriers and governed by an open standard (GSM/3GPP). SMS was never under the unilateral control of any single entity, despite many carries enjoying a defacto monopoly in their respective home markets. WhatsApp , on the other hand, is a single proprietary protocol, with a single operator, optimised increasingly for the commercial interests of that operator, and treated by the rest of the world (governments, hospitals, schools, small businesses, families separated by borders) as a public utility. The seven-day scenario above is an exercise in realising this dependency. Meta has no public-service mandate and WhatsApp ’s terms of service explicitly disclaim any commitment to availability. Yet a meaningful fraction of the medical communication, emergency coordination, family contact, and small-business activity of the global south runs on top of this disclaimed-availability infrastructure. At this point the LinkedIn thought-leadership crowd would tell you the answer is “diversification” or “resilience” or “multi-channel strategy” and add an inspirational quote alongside the ChatGPT -inserted emojis. Telling a Karachi tailor with 14 customers in a WhatsApp group to “diversify their customer-communication stack” does nothing to solve the problem. The infrastructure they depend on was built and made free at the point of use by a corporation that calculated, correctly, that owning that infrastructure was worth more than charging for it. The bill is paid in attention, in advertising, in data, and in the asymmetric power Meta now holds over a substantial fraction of global communication. While the shutdown will (sadly) not happen any time soon, the dependency, however, exists, and the thought experiment is worth running occasionally (with other services as well… looking at you, Google Mail !) because this exact dependency is what should push us to look for alternatives, and not the implausible event that would make it visible. Network effects may be the biggest drivers for this unhealthy dependency, but I believe that each and every person has the ability to make an impact within their families, their friend-circles and their communities, by choosing to use anything but WhatsApp as their main communications channel, ideally a self-hosted alternative . For almost three decades now we’ve had XMPP available to us, with popular and capable implementations like ejabberd , Prosody , and Snikket existing as open-source software that is ready to be used for communications platforms of any size. As a matter of fact, WhatsApp uses XMPP behind the scenes and is in fact built upon the same great technology stack used by ejabberd . For a “lower-level” alternative, there’s the good ol’ IRC that has been around for almost four decades and that is still thriving . Both of these open standards would allow communities, organisations and even whole governments to build public infrastructure that could in large parts replace WhatsApp . PS: Are you a Jabber user already? Come join the community channel !

media Media business Business ruby Ruby
0 views
Sean Goedecke 2 days ago

What does "playing politics" mean for software engineers?

Software engineers are often told to “start playing politics”, but most engineers have no idea what that means. Their reference point for “playing politics” comes from fiction like Game of Thrones. Are they supposed to raise an army and depose the CEO, or poison each other at team lunch? Should they book Zoom calls with each other and plot schemes? All of that is obviously ridiculous. In terms of Game of Thrones, software engineers are not lords and ladies. We’re the soldiers and workers of the realm. So you should think about “playing politics” in the way a castle guard would, not one of the major players. The castle guard are not going around poisoning people or forming coalitions between the great powers. They are largely keeping their heads down. But in order to do that, they have to stay aware of the political currents, or they’re liable to do something catastrophically stupid: for instance, making an enemy of a powerful courtier, or arresting somebody who’s on an important mission for the king. Given that, the basic principles of playing politics are something like this: As a software engineer in a large company, you will not be a powerful person . Powerful people are typically in senior management: VPs, directors, and so on 1 . However, not everyone in senior management is powerful. Some are killers who have the active support of the CEO, while others are confused incompetents. How do you know which is which? If someone is clearly ferociously competent, they’re always going to have some power, since upper management tend not to ignore useful tools. But you can’t rely on competence as your only guide. Some managers are powerful for other reasons: they’re friends with the CEO, or they have strong relationships with other groups like legal or sales, or they’re simply willing to do whatever upper management wants done. One signal is who’s leading the important projects. Read your CEO or CTO’s internal updates and pay attention to the projects that are called out by name. Organizations tend to give key tasks to trusted lieutenants. If a manager is leading an area that’s never under the spotlight , they probably don’t have enough clout. Another signal is hiring. Is a manager’s team growing or shrinking? Particularly post-ZIRP , headcount is a rare and precious resource. A manager who’s able to get it is likely a powerful manager, or at least is reporting to a powerful director or VP (which often amounts to the same thing). First, you should try not to make any enemies at all. Most software engineers who get “playing politics” wrong do it by needlessly alienating people: by being rude, unhelpful, abrasive, making non-technical people feel stupid, and so on. This post isn’t really about that. I’m assuming that you can figure out how to be a generically pleasant person on your own. However, competent software engineers will make some enemies . If you’re out there making projects happen, some people aren’t going to like the way you do it, and won’t be a fan of any compromise you offer. I wrote about this in Big tech engineers need big egos : the only way to avoid making enemies is to change nothing, but that’s incompatible with doing the job. Given that, be selective about which enemies you make. If you’re making a technical decision that’s either going to require work from team A or team B, and neither team wants to do it, you should try to pick the team with the least political cover. If you need a powerful VP’s team to do something they won’t like, try to be maximally respectful about it: get that team’s core engineers on-side if you can, or book a meeting with the powerful manager and explain the situation, or (better yet) ask the powerful manager sponsoring your project to go and talk to the other VP for you. (If you don’t have a powerful manager like this, consider abandoning your project). Give way to powerful managers when at all possible. Every so often you really do have to stand your ground — if the system will truly collapse otherwise, or a major customer will have an incident, or if the technical decision really is entirely bone-headed — but almost all cases are not like this. The best advice I’ve ever gotten about playing politics came from a manager I worked with long ago 2 : This is not the hill you want to die on. When I’m about to pick a fight or say something argumentative, and I’m not 100% convinced it’s necessary, I ask myself: is this the hill I want to die on? And it never is. The three rules about disagreeing with powerful people are: Disagreeing in private rarely hurts, if you follow these rules. In fact, it can help. If you can manage to disagree with a manager, get overruled, and then follow their plan without complaining, that can be the best way to gain a powerful friend. But if they think you’re going to keep griping about it, or worse still, complain to the rest of the team and foment some kind of rebellion, there’s no quicker way to make a powerful enemy. If you have powerful enemies at a company (for instance, the CTO or an influential VP doesn’t like you), quit . It’s really that bad. I have never seen this situation turn itself around, except in the very rare case where the CTO or VP is already looking for greener pastures and jumps ship. You cannot recover the situation: they have no incentive to give you the chance to change their mind, and they have almost unlimited ability to screw you on promotions, raises and layoffs. That’s why this piece of advice is second in the list. If you aren’t helpful or if your contributions are invisible, you can work on that and fix it. But if you’ve made powerful enemies, you’re done for. Just as it’s fatal to make powerful enemies, it’s very useful to make powerful friends. How can you do this? Remember you’re a palace guard, not a great lord: you make friends by doing your job . However, you can choose to do your job a little more proactively and diligently when you’re doing it for someone with political clout. One obvious application of this principle is that you should answer Slack messages from powerful people immediately . If you see an ordinary Slack question pop up while you’re doing some task, it’s okay to get to it when you get to it. In fact, it’s ideal not to respond to all questions immediately, so you don’t set unreasonable expectations (and so you don’t seem like you’re sitting around doing nothing). But when a VP comes in with a question, don’t make them wait: answer the question immediately. If the question requires research, send a “let me look into that right now” message, then do the research. This is the easiest way to get a reputation for being helpful 3 . Another way to do this is to lean in on important projects . Suppose you do ten projects in a year. Eight of them are normal, low-priority projects, and two of them are high-profile (say, finishing some big feature before your company’s yearly conference). It’s a mistake to allocate your effort equally to all ten. I wrote about this at length in Doing nothing at work : you should be operating at 80% capacity (or less), so you can then ramp up to 120% when it really matters. Pay attention to the narrative that powerful people are trying to push. Here are some potential narratives: You don’t necessarily have to jump in and start cheerleading, but you should at least not do anything that you know is going to make the narrative look weak. For example, on that last point, it’s foolish to openly argue that the project really was fine all along. Bring it up privately, not publicly, or you risk ruining some clever piece of propaganda that the manager in question is trying to push on the rest of the organization 4 . Finally, an underrated way to help powerful people is to offer them social support and information. Slack messages and planning emails might seem unimportant to you, but powerful people often live in that environment: their primary tool is writing messages like these, just like your primary tool is writing code. Reading and responding (in a supportive way) to these messages is something that most engineers don’t bother to do, but it goes a long way. Likewise, dropping a senior manager a line now and then (say, a heads-up that a particular project landed successfully, or that you got good metrics about some feature) is surprisingly helpful. Senior managers live in an information-poor environment: for them to learn something about a team’s work, that information has to bubble up through several layers of interpretation and summary. In my experience, they’re appreciative of being drip-fed the occasional piece of information, so long as you keep it brief and relatively rare. If you’re directly responding to a VP’s Slack messages or DMing them information, they know you’re the one doing it. But if you’re just doing your job and working hard on projects they care about, they might not notice. Being invisible is probably the most common way engineers fail at playing politics. Fortunately the fix is simple: tell people what you’re doing. If you fix an important bug for a launch, write a message in that launch’s Slack channel saying “hey, I just fixed this bug”. What if you don’t like bragging? Get over it. You have to be comfortable publicly telling people what you’ve done. You should also keep a brag document so you can repeat all of this at review time. Another, subtler way to do this is to gain the trust and respect of the powerful engineers in your area. Senior managers will always have a few trusted engineers they rely on to assess technical questions. They will ask those engineers what they think about you, and will broadly trust those answers. The good news is that if you’re competent and useful, those engineers will already value you, so you don’t have to do anything special: just be good at your job. Is playing politics all about sucking up to senior managers? Basically, yeah. A less cynical way to describe it would be “aligning with the values of the company”. If you think your company is doing good things, you should want to do that anyway! In any case, what that comes down to is figuring out what the people in charge want, giving it to them, and making sure they see you doing it. However, there’s still some scope to get what you want out of the deal. I said earlier that software engineers do not wield organizational power. However, that doesn’t mean you’re powerless. Technical ability is a source of real power, if a delicate and unreliable one. The movers and shakers in tech companies are utterly dependent on technical people to implement their vision and to give them clear answers about the system. There are many subtle ways you can leverage this. One I wrote about in How I influence tech company politics as a staff software engineer is to wait until important people at the company want to do something (say, improve reliability), then offer them a technical plan that does it your way. Another one is to become so useful that you’re actively in demand to lead projects, and then run the project how you want. You probably won’t be able to change the company’s grand strategy. But how that strategy is implemented has a lot of specific technical detail, and you can put yourself in a position to decide on those details. Playing politics isn’t about plotting and scheming, and it isn’t just about being a friendly, likeable person (although that helps). It’s about figuring out how your company actually operates: who makes the decisions, who gets consulted, what behavior gets rewarded, and so on. The most basic way to do that is to figure out who is powerful, get out of their way, and (if you can) help them get what they want . Obviously the exact titles depend on your company. One person I’m deliberately leaving out is your own manager. In general don’t think your relationship with your own manager counts as “playing politics”: that’s just you getting along with another human being. An exception to that is if you report directly to a powerful director or VP. Ironically, this manager struggled to take his own advice. Note that you actually have to be able to answer their question accurately in order to do this. If you’re not competent enough to be useful to powerful people, you will struggle to befriend them. For instance, maybe the CEO is convinced that the project was in bad shape because of something he heard, and the manager in question knows it’s easier to sell “yes, but we turned it around” than “no, you misunderstood, everything was always fine”. If you complicate that process, you risk the CEO thinking that the project is still bad and cancelling it. Be aware of who’s powerful and who’s not At all costs, avoid making powerful enemies Help powerful people as best you can Make sure they know you’re helping them (without annoying them) Make sure you do it in private When they overrule you, stop arguing immediately We’ve had a lot of turnover and reorgs lately, but we’re all starting to pull together as a team now Isn’t it great how focused we all are on reliability work after last month’s incident? The conference this week is the most important thing, so we’re all being very careful not to break anything We’re an AI-forward team that’s looking for the best ways we can leverage LLMs into our team processes Although this project had a rocky start, we’re now all aligned on the way forward Obviously the exact titles depend on your company. One person I’m deliberately leaving out is your own manager. In general don’t think your relationship with your own manager counts as “playing politics”: that’s just you getting along with another human being. An exception to that is if you report directly to a powerful director or VP. ↩ Ironically, this manager struggled to take his own advice. ↩ Note that you actually have to be able to answer their question accurately in order to do this. If you’re not competent enough to be useful to powerful people, you will struggle to befriend them. ↩ For instance, maybe the CEO is convinced that the project was in bad shape because of something he heard, and the manager in question knows it’s easier to sell “yes, but we turned it around” than “no, you misunderstood, everything was always fine”. If you complicate that process, you risk the CEO thinking that the project is still bad and cancelling it. ↩

career Career business Business
0 views
Stratechery 3 days ago

Apple Sues OpenAI, Apple’s Real Problem

Apple is suing AI for stealing trade secrets; there is one guilty employee, but this mostly feels like lashing out.

ai AI security Security business Business
0 views
DHH 4 days ago

The will to power will return

In the 1980s, France started 43 nuclear reactors across 14 sites. On average, each reactor took just seven years to build. Forty years later, all but one of these reactors are still running, and they continue to produce nearly half of France's electricity. Can you imagine France doing something like this today? Or any other country in the West for that matter? The past is a foreign country. But why is this? Why did the West lose the will to power? A popular meme would explain it as the inescapable good-times-hard-times circle: Hard times (WWII) create good men, good men create good times (Les Trente Glorieuses), good times create weak men (The End of History), weak men create hard times (now). The Fourth Turning by Strauss and Howe offers a theory for this wheel of time by tracing the last five centuries to the same four recurring phases: High, Awakening, Unraveling, Crisis. It was the good men of France's hard times who planned the country's incredible nuclear build out. This hero generation, as Strauss and Howe calls them, planted the trees of power that would provide shade for several generations to come. It seems inconceivable to expect similar bold plans and action from the current cohort of the European political establishment. But The Fourth Turning argues this was ever thus. The decline that always sets in once we enter the unraveling phase of the century (or saeculum, as the book calls it) inevitably leads to a crisis. We're on the cusp/in one of those right now. So pessism is perhaps a rational response. And yet, the night is darkest before the dawn, and the current Crisis is likely to lead to another High, if the past five centuries and Strauss and Howe's theory are any guide. If so, we should expect the next hero generation to reject this managed decline of our present turning, and once again taking up the mantle of ambition. The circle of the saeculum is both a prophecy and a roadmap. We're not supposed to live like this forever: weak, ineffectual. This too shall pass. And when it does, once the Crisis becomes another High, we'll marvel at the time wasted, but with the pity due a pathetic period of the past, not from within an eternal prison of decline. We just have to make it out of the current Crisis alive. The last one brought us a total war. Would be nice if we could get back to the High without something quite as devastating, but don't bet on it.

business Business culture Culture
0 views
DHH 4 days ago

But Y

It's no mystery to me why the Tesla Model Y is the world's best-selling car. As a total package, I could make a fair argument  that it's simply because it is the world's best car.  I'm no stranger to Teslas at this point. We've owned a Model S Plaid, the Model X we traded in on the Y, and we still have the Cyberbeast too. But as impressive as all those cars are, the Y towers above them in several key respects, but first and foremost, value. The premium all-wheel-drive white-on-white seven-seater we just got was right around $55,000. That's not exactly cheap, but it's less than half of what we spent on any of the other Teslas. It's a quarter of what we spent on the Porsche Taycan Turbo S. It's a sixth of what a new Aston Martin DBX would set you back. And, if I could just have one car, I'd pick the Y over all of them. The first thing you notice coming from earlier Tesla models is just how well-built the new Model Y is. The gigapress process that produces these new cars results in a package that feels reassuringly solid: no squeaks, no rattles, no flex. This couldn't be said about any of the earlier S and X models we had. But compared to other makes, it's not exactly revolutionary that a brand-new car feels well put together. Many other makes have managed to perfect that process over the decades. Tesla has now merely leapfrogged itself to the front of the class. But what very much is revolutionary is just how effortless owning the Y feels. It starts with entry and exit. Once you've paired your phone, you never think about keys or starting or stopping the car again. It just happens. There's no on/off button, no starter, no unlock. Again, other makes have made attempts at this, but none that I've tried is even close to the effortlessness that Tesla's superior software stack is able to deliver. Speaking of software: It just works. Every time. Going anywhere. You don't miss Apple CarPlay or Android Auto for a second. The navigation, the Spotify integration, the setup. Everything feels like it was written by a leading American software company. Not subcontractors out of India or firmware developers forced to deal with user interfaces. But where everything comes together is FSD. The self-driving technology that Tesla pushed against all odds for over a decade is finally here in an utterly magical incarnation. The car not just drives itself anywhere, it drives better than almost any human I've ever been driven by has been able to do. Its ability to anticipate traffic patterns, hit the perfect deceleration curve towards a light, slow down for even minor speed bumps, and gracefully curve around pedestrians or cyclists is nearly unbelievable.  As in, you'd be forgiven the suspicion that there must be a human driver hidden somewhere controlling the car over the internet. But it's just AI, and it's gotten fiendishly better over just the past year or so. All in service of that effortless experience. In fact, I'd go so far as to call it a luxurious experience. Like you're being escorted by the Queen's own driver to your desired destination. The Queen wouldn't bother with keys or rattles or driving. She'd just get in, be driven, and arrive fresh for a waive. This is the best approximation you can buy for mortal money today. But then, unlike the old X, it's actually also surprisingly delightful to grab the wheel yourself, hustle it down a hill, lean it into some fun corners, and surge out on that wave of endless torque that electric motors always deliver so well.  No, it's not a Porsche 911, but I'd say it's 90% as fun as a Taycan, at a fraction of the price, in a package that's endlessly more practical, and — did I mention this already? — can drive itself once you're done with the spirited part of the journey. The Tesla Model Y is a triumph of capitalism. Making the best self-driving technology available to the masses at a price that's accessible to the middle class in a car that even billionaires would appreciate.  Andy Warhol captured this egalitarian celebration well with this sentiment: “A Coke is a Coke and no amount of money can get you a better Coke than the one the bum on the corner is drinking. All the Cokes are the same and all the Cokes are good.” The Tesla Model Y is an incredible car for nearly everyone.

programming Programming Electronics business Business
0 views
Ed Zitron's Where's Your Ed At 6 days ago

Premium: The Hater's Guide To The Memory Crisis

Hi premium readers! I’ll be taking a week off of the premium next week — July 17 — to have some well-earned rest. This will mark only the second time I’ve missed a premium piece since I started this newsletter in June 2025, and I hope you’ll forgive me for the (short) break. Don’t worry. Today’s piece is also an absolute banger. Everything’s more expensive, and it’s all AI’s fault. It really is that simple.  An AI data center is full of servers, which are in turn full of (for the most part) NVIDIA GPUs. Each NVIDIA GB300 has two B300 GPUs, the two of which have 576GB of High Bandwidth Memory (HBM, or HBM3e to be specific), and a CPU, which has 480GB of lower-power LPDDR5X RAM (the kind usually used in cellphones and other mobile devices). These systems tend to be sold in an NVL72 rack with 18 compute trays, bringing us to 36 GB300s , for a total of 20.7 terabytes of HBM and 17 terabytes of LPDDR5X RAM, and that’s before you get to the RAM associated with the high-speed networking gear and other associated components. Analyst estimates have the cost of the high bandwidth memory of a single NVL72 GB300 at around $15.27 per gigabyte, for a total of around $316,000 of HBM, and while I can’t seem to find a stable source for pricing around LPDDR5X, I think a fair estimate is around $4 per gigabyte based on this piece , so around $68,000 worth per NVL72 rack. At around 150kW of power draw per NVL72 , a 1GW data center (with 740MW of critical IT load) would have around 4,933 NVL7s racks — for a total of $ 1.894 billion in HBM and LPDDR5X costs, or around $2.559 million of HBM and LPDDR5X RAM per megawatt of IT load.  Oh, and each of these NVL72s can hold as much as a petabyte of expensive solid state storage, costing an additional tens of thousands of dollars.  Because HBM takes up more space on a wafer — the slice of semiconductor material that is etched using photolithography ( read: molten tin ) and then cut into separate dies (individual chips) — and generally has much higher margins (thanks to the triopoly of Samsung, SK Hynix and Micron), memory manufacturers are dedicating more space on their manufacturing lines to it than to regular consumer RAM, which allows (thanks to said triopoly) said manufacturers to charge effectively whatever they want for consumer RAM. And thanks to AI — to quote Tom’s Hardware and Counterpoint Research — NVIDIA is buying that LPDDR5X RAM at the scale of an Apple or a Samsung: The net result is pretty simple: every single consumer electronic of any kind is getting more expensive. Valve’s Steam Machine console debuted at a 30% higher price point than planned , Apple hiked the prices of its MacBooks and iPads and will likely have to do the same for its next iPhone . Nintendo , Microsoft and Sony increased the cost of their consoles, and the PS5 and Xbox Series now cost more today than they did when they first retailed, almost six years ago.  On the Android front, Samsung has bumped the price of its Galaxy smartphones , and manufacturers in this space (which tends to have smaller margins than those enjoyed by Apple) are likely to limit the number of new devices shipping with 16GB of RAM, as well as re-introduce models with 4GB of RAM   .  Meanwhile, memory manufacturers are having record quarters, with Micron’s revenue quadrupling year-over-year in Q3 2026 and its gross margin improving by ten percent (from 74.9% to 84.9%) quarter-over-quarter, and Samsung’s profits growing from $38 billion to $59 billion quarter-over-quarter thanks to the spiralling cost of revenue caused by…well…the companies setting the price of memory at whatever they’d like. This is a problem caused by the fact that these three companies — SK Hynix, Micron and Samsung — produce more than 90% of the world’s RAM, which is why there’s a price fixing lawsuit against them , per Polygon: To be clear, HBM is more expensive to make than regular RAM, and takes up significantly more space ( about 4x more ) on the wafer, but because of the incredible demand for AI servers, Samsung, SK Hynix, and Micron can charge effectively whatever they want for it, much like they are for the regular RAM that’s in short supply. The same is becoming increasingly true for the solid state storage that these companies (and others like Sandisk) sell too. Now, you may think it’s a little rich to suggest that memory manufacturers are colluding to rig their prices, perhaps a little judgmental , and you’d be wrong because they’ve done it before. Quoting Polygon again : To be clear, I am not saying — nor can I prove — that there is any kind of price-fixing or collusion going on. Nevertheless, there are three companies that effectively make all the world’s RAM, all raising prices at the same time, all seeing record profits, all riding high at a time when everybody else is suffering as a direct result.  The Wall Street Journal put it best : What makes this particular memory crisis so distinctly dangerous is that it isn’t a result of consumer demand so much as it is capital expenditures from very large companies making bets that don’t connect with reality.   Microsoft, Google, Amazon, and Meta aren’t spending $765 billion in capex in 2026 because of rapid demand by consumers for AI services, but a desperation caused by a lack of hypergrowth ideas , circular financing with Anthropic and OpenAI , and a vague concern that if they stop spending that the other guy will do something as a result. As I discussed earlier in the week , nobody can make a compelling case for building more data centers other than “we must do so, because of AI.” Nobody is having trouble accessing ChatGPT, Claude or another major AI service because of a lack of compute, outside of Anthropic and OpenAI’s continual rapacious hunger for more compute that doesn’t ever seem to involve them turning away business. While price increases generally help moderate demand for goods or services, none of that matters when you have four companies willing to spend a trillion dollars a year on the off chance that they might get something out of it .  As a result, Micron, Samsung, and SK Hynix can charge effectively as much as they want, and NVIDIA and others building black holes for AI capex can then pass those costs onto Microsoft, Google, Amazon, and Meta, who have given themselves a blank check to build whatever it is that they think will come out of the large language model era. Put another way, the capex spend of four of the largest companies of the world — all of whom are now funding their capex using debt — has now led to the single-largest increase in the price of consumer electronics in history, for the most part thanks to one company, NVIDIA, becoming the largest purchaser of HBM in the world because those four companies are buying so many GPUs.  To give you an idea of how bad that is, NVIDIA takes up roughly 65% of all high bandwidth memory, with the other 35% (mostly) going to specialist ASICs from Google and Amazon, and AMD’s Instinct line of AI GPUs.  This is a unique — and uniquely dangerous — bubble, because demand isn’t based on actual revenues or events happening outside of those in the imaginations of Sundar Pichai, Mark Zuckerberg, Andy Jassy and Satya Nadella. They didn’t start buying these GPUs because consumers demanded them. In fact, they did so without really checking whether consumers gave a shit, which is why I’m so worried about what comes next.  Only 23% of total DRAM wafers are taken up by HBM , but it’s accounting for a remarkable chunk of revenues, at least for SK Hynix, where it took up 40% of all DRAM sales back in Q3 2025 , the most-recent number I can get.  While I can’t find definitive numbers from Samsung or Micron, the situation is bad no matter which way you spin it. Either they’re increasingly-relying on HBM as a revenue driver to the point it’s crowding out the revenue from their other DRAM businesses (making them dependent on GPU and ASIC revenue), or their revenues are spiking because they’re able to crank up the cost of DRAM. This is setting everybody up for a dramatic and painful collapse, largely based on the strange nature of how memory is built and sold, unless cooler heads prevail and capex doesn’t accelerate based on hopium.  What happens when hyperscalers reduce their capex, or when banks stop issuing data center debt ? NVIDIA stops needing all that HBM, which means any and all capex dedicated to expanding manufacturing  infrastructure to produce more HBM — which is not particularly valuable outside of AI GPUs — will have been built to capture demand that doesn’t exist. While that capacity could be re-engineered to make useful DRAM with mass appeal, doing so will also drag down the profits of every memory manufacturer in the process, creating a supply glut the likes of which we’ve never seen in history.  The memory industry has gambled its financial future on the idea that there’s near-infinite amounts of capital available for data center capex, adjusting its supply chains and fabs to focus on scooping up demand that’s increasingly only made possible by the availability of debt. Microsoft, Google, Amazon and Meta have turned NVIDIA into a single point of failure for the entire tech industry, creating a painful present for consumers and a brutal future for suppliers, all because they decided to spend more than a trillion dollars on a dead end industry. The longer it takes for hyperscaler capex to retract, the more expensive everything becomes. The more GPUs that get sold, the more capacity that gets put toward high bandwidth memory, and the more that Micron, SK Hynix and Samsung can charge for it, which makes it more expensive to buy AI GPUs, which increases the amount that hyperscalers are spending on AI capex for effectively the same amount of gear. The longer that hyperscalers sustain this pace, the larger the return needs to be, and at this point, none of them have disclosed their AI revenues, which heavily suggests there’s yet to be a dollar of profit.  Yet the more they commit, the more committed they have to be. Pulling back at this point will prove to the markets that they’ve committed to too much capacity. Yet not pulling back means that hyperscalers will continue to turn their free cash flows negative in pursuit of an indeterminate goal. It’s a vicious cycle made worse by the fact that every spin of the capex wheel increases the price of just about every consumer electronic in the world , creating a market-wide inflation for what amounts to a speculative asset bubble. And If even one hyperscaler cuts their capex, the cartel-like memory industry is in for a nightmare scenario, one larger and uglier than any they’ve ever faced.  In the end, it all comes down to whose problem this high bandwidth memory becomes. Will SK Hynix, Samsung, and Micron have already built the RAM and face waves of cancellations, resulting in a bunch of fallow inventory it can’t use or sell? Or will they already have shipped it off to NVIDIA and ASIC builders, only for it to sit in warehouses waiting for the day it can finally be melted down? Who will end up holding the bag? The cartel of horrible fab-gargoyles, Jensen Huang’s Wallet Inspection Firm, one of the four simpleton hyperscalers, Broadcom, or one of the Taiwanese ODMs?  Just to be clear: everybody loses, unless the AI bubble continues in perpetuity. This is the Hater’s Guide To The Memory Crisis — and the terrible tale of the boom-and-bust memory industry.

ai AI business Business hardware Hardware
0 views
Stratechery 6 days ago

2026.28: XBOX On the Rocks

Welcome back to This Week in Stratechery! As a reminder, each week, every Friday, we’re sending out this overview of content in the Stratechery bundle; highlighted links are free for everyone . Additionally, you have complete control over what we send to you. If you don’t want to receive This Week in Stratechery emails (there is no podcast), please uncheck the box in your delivery settings . On that note, here were a few of our favorites this week. This week’s Asianometry video is on TOTO: From Toilets to E-Chucks . A Word from Mark Zuckerberg*.  I was delighted to see Ben insert himself into the CEO chair at Meta on Tuesday and write a script for Mark Zuckerberg as he tells the story of Meta and its AI investments in 2026. That article traces past Meta mistakes as well as those of investors who doubted the company, all to frame current investments in AI and the massive opportunities that remain central to the Meta’s future. A combination of history, analysis of the future, and fun, it’s a perfect summer read. As for a summer listen, we doubled back on all of it, plus Meta’s Muse-Spark release, for this week’s episode of Sharp Tech .  — Andrew Sharp Pulling the Plug on XBOX? It’s been years since there was good news coming out of the XBOX division at Microsoft and that trend continued this week, as XBOX CEO Asha Sharma announced plans to eliminate 3,200 jobs, or around 20% of its staff over the next 12 months. Wednesday’s Daily Update explores how Microsoft arrived at this point and why, in particular, the Game Pass initiative that was the last great hope for XBOX has been a failure. I’m not a gamer, but Ben’s rendering of the XBOX story — and the Game Pass story — is a great case study of both internet economics and management mistakes (and analyst ones!). — AS Toilet Talk . Look, I get that’s a little weird, but if there is one brand of household appliances that I cannot imagine living without, it is in the bathroom. Specifically, I absolutely love my Toto toilet, and was delighted that Jon made a video about the company on Asianometry . Here’s the twist: the reason why Toto is a subject of interest isn’t their toilets, but rather the fact the Japanese company also plays a critical role in the AI supply chain. — Ben Thompson A Script for Mark Zuckerberg — A script for what Mark Zuckerberg should say on Meta’s next earnings call. XBOX Cuts; Bundling and the Internet Solvent; Transaction, Coordination, and Sunk Costs — Microsoft’s Xbox division is conducting big layoffs, as the company deals with abject failure of its Game Pass strategy. Muse Image, Grok 4.5, Alex Karp on CNBC — The battle for verifiable data is increasingly defining the AI race, from Meta to Grok to the frontier labs. Online Insanity and Its Counterpoint — What we can and can’t achieve in response to paranoia and extremism online. The New ChatGPT App The Debt-Fueled Collapse of China’s Top Machine Tool Maker RCA and the Vacuum Tube’s Last Stand A Missile Test and New PLA Generals; The CITIC Plane Crash; America’s Taiwan Interests; Guo Wengui Jailed and Ezra Jin Released A Tale of Two Cities and Jaylen Brown, Minnesota’s Bet on LaMelo, Peterson Arrives and Mitchell Cashes Out Meta and Its Messaging Problem, The XBOX Reset, Q&A on Token Costs, American Soccer, Starlink in Nature

gaming Gaming business Business
0 views
Jim Nielsen 1 weeks ago

Family Feud: Mac-assed Mac App Edition

“We asked 100 people: What are the top three companies on earth best positioned to make a world-class Mac-assed Mac app ?” Survey says: Yes! Apple at the number one spot. Makes sense. Who better to make the very definition of a great Mac app than the people who make the Mac? No brainer, I suppose. Granted, they’ve had some misses , but nobody bats 1000. Ok, let’s keep going. “We asked 100 people: What are the top three companies on earth best positioned to make a world-class Mac-assed Mac app?” “Anthropic!” Survey says: Wow, that’s odd huh? You’d think Anthropic would be right there at number two. Not only do they have billions of dollars, but they also develop, maintain, and control the super intelligence we’ll all soon be subservient too, right? Surely if anyone (besides Apple) is well positioned to make a world-class Mac app, it would have to be Anthropic — right? And yet, here we are with Claude Desktop as an Electron app . Ok, let’s keep going. “We asked 100 people: What are the top three companies on earth best positioned to make a world-class Mac-assed Mac app?” Maybe not . Not so much . I’m sorry, but that’s three strikes. Apparently it’s a mistake to assume that a big company with piles of cash is well poised to make a great Mac app — even if they are enabled by hyper-super-intelligence. “Well who cares? It just goes to show you don’t have to make a good Mac app to be obscenely successful in terms of revenue!” Well, maybe that’s true. Actually, come to think of it, it kinda does seem like the bigger you get and the more money you make, the more likely it is you’re making an Electron app. There seems to be a correlation between “Mac-assed Mac app-edness” and “Company size/revenue”. Why is that? I’ll leave that as an exercise for the reader (though my mind is leaning towards something to do with care ). Thank you for playing reading this game of family feud. Reply via: Email · Mastodon · Bluesky

business Business
0 views
Krebs on Security 1 weeks ago

Felons, Fraudsters Flog Offensive Cybersecurity Startup

A cybersecurity startup dangling millions of dollars to acquire zero-day security vulnerabilities in popular software is run by a pair of far-right conspiracy theorists and convicted felons whose most recent ventures included fake intelligence companies and a now-defunct AI-based lobbying platform they operated under assumed names. The X/Twitter account IRIS C2 (@C2IRIS) has gained more than 4,000 followers since its creation in January 2025, posting frequently about security vulnerabilities, AI and software exploits. IRIS C2 says it is a company in McLean, Va. that sells offensive cybersecurity capabilities. The IRIS C2 website dangles the possibility of million-dollar payouts for exploits to attract talent. “Our business model is this,” reads a pinned post on top of the IRIS C2 account on X. “Attract the very best vulnerability researchers and exploit developers in the world to join our company. This mostly revolves around junior engineers with raw talent/extremely high IQ. We don’t care if they have a college degree/industry experience.” The website linked in that profile — irisc2[.]com — says the company is hiring for a number of open positions, and a recent post on its LinkedIn page enthuses about an overwhelming number of applications from potential employees. The website claims IRIS C2 is in the business of acquiring “zero-day exploits, individual primitives, partial chains, and full capabilities across all major platforms. Payouts range from $10,000 to $7 million depending on target, reliability, and operational value.” The government contracting portal g2exchange.com reports that irisc2[.]com is operated by a business based in Virginia called Calvexa Group LLC . The “contact” link on the website for Calvexa Group — calvexagroup[.]com — forwards visitors to irisc2[.]com. G2Exchange shows that while Calvexa Group LLC is registered as a federal contractor, it does not appear to be working on any direct government contracts. A search on the Arlington, Va. address listed in the incorporation records for Calvexa Group LLC finds the property is occupied by Jack Burkman , the 60-year-old founder and managing partner of the lobbying firm Burkman & Associates . When approached with questions about IRIS C2, Burkman referred further inquiries to his longtime associate, 28-year-old Jacob Wohl . Jack Burkman (left) and Jacob Wohl, at a press conference in August 2020. Image: Wikipedia. Burkman and Wohl have a storied history of creating fake intelligence companies and using them to spread false claims about and frame public figures, including fabricated sexual assault claims against then FBI director Robert Mueller , and Pete Buttigieg , then mayor of South Bend, Indiana and a Democratic candidate for the presidency. In 2019, Burkman and Wohl held press conferences falsely alleging extramarital affairs by Sen. Elizabeth Warren (D-Mass.) and then-2020 presidential candidate Kamala Harris . In the wake of the 2020 presidential election, Wohl and Burkman were prosecuted by multiple U.S. states for making thousands of robocalls to residents of battleground states and disseminating false claims about mail-in ballots. They were indicted in Cleveland on 15 felony counts of orchestrating a robocall scheme aimed at suppressing the black vote in Detroit, and were sentenced in late 2025 to probation after their appeals to dismiss the charges were rejected. In 2022, Wohl and Burkman both pleaded guilty to a single felony charge of telecommunications fraud in Ohio, and sentenced to a fine, probation, and community service. In March 2023, a judge in a New York civil case ruled that Wohl and Burkman had violated federal and state civil rights laws, and the two agreed to pay a $1 million settlement. In June 2023, the Federal Communications Commission (FCC) imposed a $5.1 million fine against Wohl and Burkman for their robocall campaigns, at the time the largest fine ever sought by the FCC under the Telephone Consumer Protection Act. Jacob “Jay” Wohl’s GitHub account. By the age of 17, Wohl had started multiple investment firms, and cultivated the nickname “Wohl of Wall Street” after appearing on Fox News in 2015 to discuss his new hedge funds. In 2017, the Arizona Corporation Commission charged Wohl and his investment funds with 14 counts of securities fraud, and ordered him to pay $35,000 in restitution. In 2019, Wohl pleaded guilty in California to four felony counts of selling unregistered securities and was sentenced to two years of probation. The market for previously unknown security vulnerabilities has always been populated by a colorful mix of researchers, academics, charlatans, clout-chasers and people actively involved in cybercrime communities. But the market for selling offensive security services to the U.S. government tends to be far more circumspect. Plenty of government contractors recruit vulnerability researchers and pay for the exclusive rights to novel software exploits, yet none of them do so quite as brazenly and openly as IRIS C2. Recent posts from the Twitter/X account IRISC2 (@c2iris). Indeed, KrebsOnSecurity was unaware of IRIS C2 until last month, when an attendee at a regional cybersecurity conference shared that Wohl and Calvexa Group were pestering people at the conference about selling their vulnerability research. In an interview with KrebsOnSecurity, Wohl said Mr. Burkman was not involved in the day-to-day operations of IRIS C2. Wohl shared that IRIS C2 originally began as a penetration testing company, but shifted its focus recently to selling phone-hacking services to the government. Several times throughout the interview, Mr. Wohl mentioned working on federal government contracts, but when pressed for specifics said he was not at liberty to speak publicly about them. Mr. Wohl said he does not have any formal education or training in computer science or information security, and that most of his knowledge on the matter is self-taught. “I know more about tech than anyone,” Wohl bragged. “My background has always been extremely technical, and I’ve always been deeply into tech. People know me as someone who is able to create spectacularly exquisite capabilities that would make your head spin.” Wohl said security researchers bring the company unique vulnerability findings “on a regular basis,” but that in many cases those findings are preliminary and not fully fleshed-out. “Let’s say someone finds a flaw in a media decoder on a phone,” Wohl said. “A lot of times what we receive is an exploit primitive, where the idea is there but the [execution] needs work. You need that exploit to be stable and reliable, and that’s what we do.” Wohl claims IRIS C2 has approximately 40 employees, although he said none of them are allowed to list their employment on LinkedIn for operational security reasons. In May, the author of the IRIS C2 account on X said that his girlfriend had no idea what he did for a living. But if IRIS C2 has any other employees, they may be similarly unaware of Mr. Wohl’s history of outright fabrications — or even his real name. In September 2024, Politico reported that Burkman and Wohl were bragging about big companies supposedly buying services from their now-defunct company LobbyMatic , which claimed to use artificial intelligence to assist in political lobbying efforts. However, Politico found the pair were running the company using pseudonyms, with Wohl reportedly adopting the name “Jay Klein” and Burkman using the moniker “Bill Sanders.” Politico reported that two of the former LobbyMatic employees resigned after learning of their true identities, while other employees only learned after they had left the company.

Crime business Business security Security
0 views
Stratechery 1 weeks ago

XBOX Cuts; Bundling and the Internet Solvent; Transaction, Coordination, and Sunk Costs

Microsoft's Xbox division is conducting big layoffs, as the company deals with abject failure of its Game Pass strategy.

business Business gaming Gaming
0 views
All Things Distributed 1 weeks ago

Now Go Build CTO Fellowship: Season 2

Today, we're releasing the second season of the Now Go Build documentary series. Five episodes featuring technology leaders from around the world solving the hardest problems in healthcare and education.

programming Programming business Business career Career go Go
0 views
Stratechery 1 weeks ago

A Script for Mark Zuckerberg

Listen to this post : The setting: Meta’s earnings call in early August, 2026. The speaker: Meta CEO Mark Zuckerberg . Good afternoon everyone, and welcome to Meta Platforms’ Second Quarter 2026 Earnings Conference Call. Our remarks today will include forward-looking statements, which are based on assumptions as of today. Actual results may differ materially as a result of various factors, including those set forth in today’s earnings press release and in our quarterly report on Form 10-Q filed with the SEC. We undertake no obligation to update any forward-looking statement. I know it’s weird that I, Mark Zuckerberg, am doing the Director of Investor Relations job, but anything is possible when this speech is made up. What follows isn’t actually me: it’s what Ben Thompson of Stratechery thinks I should say on this call. I know that Meta and myself are facing a lot of questions about AI, particularly the amount of money we are spending on capex. Our core business is an asset-light cash generation machine, so why are we spending tens of billions of dollars on AI? To answer this question I want to give you a quick recount of our history, what I’ve learned, and why I am so confident that we are doing the right thing for our future. So let’s get to it. Facebook was, as you know, the digital representation of Harvard’s analog Face Books. What was clear from the very first day we went live was the extent to which humans are, first and foremost, interested in other humans. People would spend hours clicking around to people’s pages. To put it another way, our first algorithm was human curiosity. What truly super-charged Facebook usage, however — and which transformed the Internet — was the feed. Now, instead of actively surfing to friends’ pages to look for an update, we showed updates to you in a single feed on your homepage. You might remember that we got a lot of heat for this decision, including protestors outside our office in Palo Alto. The lesson we took from that, however, is one that has guided us to this day: first, the revealed preference of users, as captured by data, was that they loved the feed: engagement skyrocketed. Second, we learned to trust our own — my own — product intuition, and that conviction has served us well over the years. Another critical moment in our early history was the shift to mobile. We didn’t get this right in the beginning — more on that in a moment — but what was quickly apparent is that more access to Facebook meant more usage of Facebook. I can’t emphasize this point enough: when humans can connect to humans, they do, and when they can do it more conveniently and in more places, they do it more often. Finally, I would be remiss to not mention Instagram. Obviously Instagram has been a major part of our growth over the last 15 years — and, I would add, we have been a major part of Instagram’s growth. To that end, an important thing to understand about Instagram is the extent to which it has evolved . Just because we gave our users what they wanted at one particular moment in time does not mean we can afford to sit still: more bandwidth first meant more pictures in Stories, and then video in Reels. Instagram has gone from strength-to-strength precisely because it has changed as technology has changed. We — I — haven’t done everything perfectly. We’ve taken our arrows through the years for lots of things that frankly aren’t our fault, but are rather the reality of being the primary communications platform for all of humanity, and humanity is flawed. I’m proud of the efforts we have made to ameliorate humanity’s worst impulses while enabling some of our best tendencies, including that desire to connect. Rather, my mistake is itself a very human one: for many years I have resisted embracing what Facebook — now Meta — is, and spent too much time trying to emulate some of the tech titans who came before me. Specifically, I have been obsessed with becoming a platform. The first manifestation of this error was the initial shift to mobile I referenced above. When Facebook was primarily a browser app I invested heavily in trying to build a platform, with things like Facebook Games, payments, etc. We had some success there — some of you on this call might have played Farmville back in the day — but when mobile came along we mistakenly tried to hold onto web technologies that supported my vision, and were years too late in investing in a truly native smartphone experience. The reality — and this is hard for me to admit — is that Apple saved us from my mistaken obsession. Mobile Made Facebook Just an App, and that was Great News . Instead of diminishing the Facebook experience so that we could feature third-party developers, we had to cede that space to Apple and put our own content front-and-center. It turns out that was what people wanted the most; in fact, they wanted it so much that they willingly scrolled through and clicked on the most compelling ad units ever. And make no mistake, we paid back our debt: Facebook built the App Store just as much as Apple did. My second error was Reality Labs. While in recent years I have framed our acquisition of Oculus and virtual reality as a necessary response to Apple’s attempt to handicap our business, the truth is that I invested twelve figures into this technology because I thought it was cool, and yes, because I wanted to own a platform. I do think we’ve made compelling strides in this area — and we’ve created technology that is going to matter in the long run — but I now recognize that part of the reason I am delivering this mea culpa right now is because I burned a lot of credibility with investors with all of the losses Reality Labs has endured with very little to show for it. My third error was not in trying to make Facebook something it was not, but rather failing to appreciate what it had become. While I was thinking about platforms, I took it for granted that connection was enough for the core business; in fact, Facebook had evolved into entertainment , at least in its public-facing forms (I will take credit for the acquisition of WhatsApp and realizing that Messaging Was Mobile’s Killer App ). This was an insight that TikTok figured out first , and it was a blindspot for me . What I’ve come to realize is that all of these mistakes are symptoms of what has been my biggest failing as CEO: all of you on this call have appreciated our ad business more than I have. I’ve been very blessed as CEO to have excellent co-workers who have over the years developed the world’s best digital ad business, while I frankly haven’t taken as much interest as I should have. My failure to appreciate our ad business is another lens through which to examine my mistakes: This neglect as CEO left us badly exposed in our disputes with Apple. I firmly believe that Apple’s characterization of digital advertising was unfair, dishonest, and self-serving . What I failed to do, not just in that bruising battle but in the years leading up to it, was make the affirmative case for ads generally, and Meta ads in particular. It’s easy to see how the Internet has made it possible for an entirely new category of entrepreneurs to create products that uniquely serve the tremendous capacity of humans to manufacture an infinite array of desires, growing the economy to the benefit of everyone; what’s harder to appreciate — in part because I haven’t made the case — is that the only way to connect those creators to the consumers who love them is digital advertising. We don’t serve ads like Google — or Apple in the App Store, or Amazon on Amazon.com — that in many respects function as a tax on search; we show people products they never knew existed, but that immediately generate desire and, ultimately, happiness. In short, I believe that we are a force for good in the world, not just because we connect people to each other, but because we connect entrepreneurs with customers in a way no one else does. Forgive the long preamble, but this is necessary context for me to properly explain why AI is so important to Meta, and why I am making the right choice to invest so heavily in both talent and infrastructure. First, when investors compliment our asset-light business, what they are complimenting is the fact that our business is purely digital. Everything digital, however, is firmly within AI’s cross-hairs. It may seem odd to begin my AI pitch by highlighting terminal value risk, but today is about honesty: every single digital company on earth faces an existential threat from AI, and we are no exception. Meta must invest in AI because a failure to do so would cost us far more in the fullness of time, particularly now that we’ve seen the very real risks entailed in depending on a third-party . Second, AI makes our business better — and by “our business”, I mean ads. AI is more than LLMs: it is machine learning, and we have been using machine learning to improve our ads business for years. More recently, we have developed GPU-dependent algorithms that have significantly improved our ability to not just target ads but also recommend content, which keeps people entertained longer, which lets us serve them more ads. And, looking forward, LLMs themselves will transform advertising, not just by generating copy and images, but by predicting the ads and content that people want to see. Every single one of these improvements goes directly to our top line — and remember, because advertising enables us to offer our products for free, the capacity to increase our top line is unbounded by price elasticity. Third, the single most important indicator that our business is on the verge of a step-change in growth is when we dramatically increase inventory. This is something investors regularly get wrong: back when we added Stories, investors panicked about falling prices-per-ad without realizing we were increasing inventory we could grow into. Five years later, investors made the exact same mistake with Reels . Those were the two best opportunities to buy Meta stock — or any stock, really — in history. We are facing an even larger opportunity over the next several years. AI makes every pixel monetizable , which means we are looking at the largest inventory expansion ever. Yes, it will take a few years to realize this opportunity, but the technology is there. More importantly, what I’ve come to realize as I’ve embraced our status as an entertainment provider and ad purveyor is that — our nature as a digital business notwithstanding — we are remarkably well-placed to thrive in an AI era. Remember what we learned about humans: they are obsessed with other humans, and they want to connect with them; that obsession and desire are only going to increase as we interact more and more with AI. AI is going to make our properties more essential, not less. Moreover — and here I must issue one more mea culpa — AI is a productivity tool, but productivity is not the end-all-be-all of the human experience. I have talked over the last year about building superintelligence that helps you get things done, but that’s a business story. What we can uniquely do is give people the experiences they want — from connection to entertainment to shopping — when they are off the clock. The fact that we are investing in AI but not selling solutions to businesses is actually one of our biggest advantages . Oh, and by the way, AI might actually lead to new hardware paradigms. I admit I was wrong to spend so much time on virtual reality, but that did lay the groundwork for a unique opportunity to develop devices that make much more sense in a world where we want to access AI everywhere, not just on a phone in our pocket. I know that many of you on this call have doubted my investment decisions before — and I understand the consternation about Reality Labs in particular. However, keep in mind that when our stock dipped in 2022, one of the big reasons was because of our aggressive capex spending, which went primarily to GPUs; ChatGPT came out a month later, and that decision to spend heavily with Nvidia looked incredibly prescient in hindsight. That prescience, however, pales in comparison to the payoff that will accrue to anyone with the foresight to build data centers and buy compute over the last several years, and for years into the future. We don’t have the luxury of waiting until the future is invented and then investing; we need to invest now, especially when the opportunity in front of us — with ads specifically — is so apparent. That noted, we are in a truly unique time, when there is a real market for selling compute on the spot market. To that end, we are going to sell access to a portion of our compute infrastructure on a short-term basis, with the ability to claw back that compute at any time. This will accomplish two important things: first, the proceeds from these rentals will fund an even larger build-out going forward. We will need this capacity in the future. Second, rental prices will provide a hurdle rate that will focus and discipline our decision-making. Let me expound on this point, because it brings this entire opening statement together: I now realize that my obsession with platforms and productivity has frequently led us astray, and that I have given insufficient appreciation to our advertising business and failed to embrace the reality of what Meta is. I truly believe we have compelling reasons to invest in AI — arguably the most compelling reasons — and the fact that the market doesn’t agree is my failure. To that end, making our compute available for rent means that we can only take it back if we can make more money on it ourselves; the only way we can do that is by leaning into what we are good at, not what I have spent too long wanting us to be. To put it another way, our best product decisions have been intuition validated by data and revealed preference; that’s how we’re going to approach AI. We will build, because we must, but we will let the market decide who gets to use it: I’m confident my newfound religion on ads will result in all of that compute being used by us to make more money than we can ever make as a permanent cloud provider. We are not out here to make chatbots or compete with OpenAI and Anthropic; they can fight for work and productivity and charging subscriptions and replacing humans. Our goal is to celebrate humans, to connect them, to entertain them, and to enable commerce among them. We need compute to do this at scale, and I know it will pay off. My commitment to you is that we will structure our business so we have no choice but to do just that. We’ve done it before, and we will do it again. And with that, over to Susan. Building a platform is antithetical to building an ad business. A platform’s goal is to feature third-parties; an advertiser’s goal is to capture attention for itself. Investing in an entirely new technology, including developing hardware, fundamentally limits our addressable market ; an advertiser’s goal is to maximize its market size. Entertainment is the best possible category for an advertiser to own: people willingly give entertainment their attention, which is exactly what an advertiser wants to sell.

ai AI programming Programming business Business
0 views
Ed Zitron's Where's Your Ed At 1 weeks ago

Premium: The Hater's Guide To SoftBank

Soundtrack: Ozzy Osbourne — Mr. Crowley A lot of people have been making a lot of fun of the SoftBank 46th annual shareholder meeting and Masayoshi Son’s (to quote Bryce Elder of the Financial Times) Untethered Goose Game , specifically referring to slides that, well, looked like this: As funny and silly as these slides might be, they’re actually very indicative of the mindset behind SoftBank. Each one of those golden eggs refers to a trillion yen (about $6.15 billion) in the Net Asset Value (NAV) of SoftBank’s holdings, with the minus referring to its debt.  It’s actually very simple, especially if you know anything about geese.  SoftBank is the goose. Masayoshi Son is the gander. Masayoshi Son mounts and impregnates SoftBank — by which I mean invests money in companies using SoftBank’s funds — at which point the goose (SoftBank) becomes pregnant (the portfolio company grows larger) and then lays the egg (the portfolio company goes public). Basically, SoftBank is a company that invests in companies that then go public and make SoftBank money, at least in theory. To continue mounting the geese , SoftBank takes on a constant flow of debt either by raising it via the bond market, taking margin loans out using its shares in successful investments like ARM or Alibaba as collateral, or (in times of trouble) outright selling shares in companies like T-Mobile or NVIDIA .  Softbank has around $50.5 billion worth of outstanding notes as of writing this sentence, not including other forms of debt, like commercial paper and traditional loans. Including those brings the total to an astonishing $76.431 billion. And, again, this is just the Softbank Group – and not any of the other affiliated entities, who have their own balance sheets and separate reporting. When Masayoshi Son protests that the “goose was not valued,” he’s saying that SoftBank isn’t given its dues for “laying golden eggs,” because the NAV of the company does not give any value to the goose that lays the golden eggs, largely because net asset value refers to the holdings of a fucking company Masayoshi, what are you talking about? Masayoshi Son’s desperate plea that “what matters is not the eggs, but the goose itself, and its power to keep laying eggs” exists to try and distract from the fact that he’s been pretty bad at fucking the goose for the last decade or so. The vast majority of SoftBank’s Net Asset Value — which is ¥48.2 trillion rather than ¥74 trillion yen, by the way! — comes from its shares in chip company ARM (¥19.15), SoftBank Vision Fund 1, (¥3.38) and SoftBank Vision Fund 2 (¥17.19). These are two venture capital funds: one very successful (VF1 includes big hits like DoorDash and ByteDance ), and one tremendously awful (VF2 includes massive losses on WeWork and Karterra ).  His one saving grace, at least on paper, is his early investments in OpenAI, turning around $64 billion (assuming it completes all $30 billion of its 2026 commitments) into a theoretical $100 billion or more, at least if OpenAI goes public, which is almost certain to- Wait, what was that? OpenAI is leaning toward IPOing in 2027 ? It hasn’t even held pre-IPO investor meetings or set a timeline ? That’s not good at all! The SoftBank Goose Engine only functions if the goose — which was not valued by the way! — continues to lay golden eggs, and in this case, the golden egg is OpenAI, and said egg is still in SoftBank’s ovary !  The problem here is that while SoftBank’s OpenAI stock is “worth $100 billion,” private stock is valued very, very differently to a public stock that you could dump on the market. This is in part because the valuations of private companies are continually overinflated by over-eager investors who, just throwing it out there, might have valued the company based on a belief that they were put on this Earth to create superintelligence rather than whether it was a good business that would continue to grow.  Per the New York Times , OpenAI’s hesitancy to go public came from a concern that it wouldn’t get a value of a trillion dollars — a worrying bit of information considering its was last valued at $765 billion, meaning that advisers were unable to make a convincing case for a listing at a meager 30% premium. This is likely why SoftBank was unable to get a $6 billion margin loan with the entirety of its OpenAI holdings as collateral . Apparently a 6% loan-to-value was too adventurous when it came to stock in what is meant to be the world’s most important company, unless, of course, it isn’t, it won’t be, and its stock is worth fuck all.  Renewed talks for a $10 billion OpenAI-backed margin loan include a guaranteed repayment of the loan if the collateral isn’t able to replace the lost funds, the kind of thing you have to say when the underlying stock ain’t worth nothin’. OpenAI is Masayoshi Son’s final gambit, as the rest of his endless gambles have gone tits-up at an historic pace. While early bets — like his $20 million investment (around $39 million in today’s money) in Alibaba turning into holdings of over $100 billion ( with all of its stock now sold ) — have floated the company for years and helped SoftBank recover from the horrors of its dot com bubble collapse,  SoftBank is now horrendously overleveraged across the board, with 85% of its ARM shares and 70% of its SoftBank Corporation tied up in loans, its entire stakes in Alibaba, T-Mobile and NVIDIA liquidated, and the vast majority of its NAV sitting in the deteriorating value of its Vision Fund 1 and its non-OpenAI Vision Fund 2 holdings. You see, SoftBank is a holding company. It does not have “revenues” or “cashflows” in the traditional sense outside of when it’s able to either sell the things it has or raise debt. As Kakashii put it , Masayoshi Son is a perpetual gambler living in an eternal boom-and-bust cycle, going from losing 96% of his paper wealth after the dot-com bubble burst to sitting at the top of a company with a $200 billion market cap and with golden eggs that are worth, on paper, hundreds of billions of dollars more.  And he’s never, ever gambled more than he has on OpenAI and the greater AI bubble. While SoftBank’s WeWork washout lost it $16 billion , SoftBank has committed or invested over $60 billion in OpenAI, as well as billions more in related counterprojects like a still-pending 75 billion Euro investment in data centers , its $4 billion acquisition of data center firm DigitalBridge , its $1 billion investment in subsidiary SB Energy to build out more data centers , and its planned $3 billion investment in overhauling a Foxconn plant in Lordstown Ohio .  The future of SoftBank relies on both OpenAI’s ability to go public and maintain a high stock price, as any public offering will likely lead to SoftBank immediately looking for a margin loan. To make matters worse, SoftBank’s other bets hinge upon the continued success of the AI industry, which hinge both on the continued success of OpenAI and there being such incredible demand for AI services (in the hundreds of billions of dollars annually). And while the geese might have been a clue, SoftBank is a very, very weird company, and the only thing weirder than SoftBank is Masayoshi Son himself. Yet as goofy and whimsical as this all might seem, SoftBank is also one of the largest companies on the Japanese stock market , valued entirely based on the value of all those golden eggs, and no matter how much value Masayoshi Son might claim his “egg factory” might have, SoftBank’s continued existence relies on its ability to increase its NAV and acquire more debt. My concerns around SoftBank were well-summarized by The Economist back in May : It’s unclear what the future looks like for SoftBank. While death is unlikely given its near-systemic presence in the Japanese economy, its continued existence at its current scale is only made possible as long as the world’s most well-funded gambler can keep his seat at the table. While it’s seen boom and bust cycles in the past, SoftBank has never been this levered, and never gambled so hard on a single entity’s success .  While this is technically a company , SoftBank exists and operates at the whim of a man with questionable idols, insane ideas, and fantastical thinking. At one point during the Dot Com Bubble, Masayoshi Son’s net worth was higher than Bill Gates ’, rising by more than $10 billion a week, before the majority of his net worth in the space of a year and sending SoftBank’s share price crashing by 93%.  Yet even when adjusted for inflation, SoftBank only invested around $2.93 billion ($1.5 billion at the time) in the heights of Dot Com mania , and spread those investments out over multiple startups. Today I’m bringing you a guide to one of the silliest companies ever founded, helmed by one of the goofiest men alive, run in a constant state of brittle leverage.  SoftBank only avoided the void in 2023 by dumping its Alibaba shares , and this time around, Masayoshi Son may have gambled too much, putting all of his eggs in one Altman-shaped basket. Welcome to the Hater’s Guide To SoftBank, or Is Masayoshi Son’s Goose Cooked?

Finance business Business
0 views
André Arko 1 weeks ago

Meet Spinel

Now that I’ve been working on Spinel Cooperative for a full year already, I have finally managed to write the very first post I meant to write about the entire situation: What is Spinel? Who is Spinel? Why is Spinel? Read on to finally receive the answers to all these and questions and more. Head over to the Spinel Cooperative blog to read Meet Spinel .

business Business career Career programming Programming
0 views
iDiallo 1 weeks ago

Amazon Basics, but for intellectual property.

Amazon has been accused several times for ripping off merchants on its platform. And every single time they denied any wrongdoing. A merchant, or anyone really, can create a product (or source it from China), then resell it on amazon. Amazon is the service provider, and hosts all the metrics concerning the products. If Amazon themselves were in the business of creating and selling products, then that creates a potential of conflict of interest. Because they have the data of all products that sell and sell well. They could replicate that success without doing any further research since the merchant has already confirmed the existence of demand. It's not surprising that Amazon Basics quickly became the best selling "private-label brand" on Amazon. They already know what sells because they have access to the data. Yet they continued to deny it, and state that they only ever use publicly available data from sellers . An Amazon spokesperson said the company believes the allegations are "factually incorrect and unsubstantiated," adding that Amazon strictly prohibits the "use or sharing of non-public, seller-specific data for the benefit of any seller, including sellers of private brands." Yet the results are right there for all to see . If you sell any product through Amazon, you are exposing your company's operations to them. If you want to keep that information to yourself, then you don't get to reach your customers, which in reality are Amazon's customers . If you want to buy something online, and get it shipped as quickly as possible, then Amazon is a blessing. Most often than not, you are not buying the product directly from Amazon. An independent store or vendor with a presence on Amazon will fulfill your order. The seller only has minor identifying characteristics on the platform. On the search result page, the space designated to the seller is small and insignificant. The customer has very few reminders that products are offered by anyone but Amazon. (Although if you want to dispute a sale, you are starkly reminded that the item is from a 3rd party vendor.) So there is no surprise when companies embrace AI internally, they are putting themselves at the risk of sharing their product with their competitors. Maybe the most obvious example is when Antropic came up with Claude Design. A tool to help users generate designs, wireframes, etc. Kinda like Figma. That's not a problem on its own, but when Antropic's chief product officer sits on Figma's board of directors, you can't say that there isn't a conflict of interest there. In fact, the chief product officer resigned from the board merely days before Claude Design was announced. He basically extracted all value from Figma then resigned. Figma's AI features are built on top of Claude. So Anthropic literally pulled an Amazon Basic on Figma . When companies force their own employees to use AI to do their day to day work, they are basically asking employees to upload company data to a 3rd party that may become a competitor. Sure something in the contract clause says that the AI company won't train on enterprise customer data, but nothing stops them from peaking at successful product data. Whenever someone tells me that they used AI to build an app and boast of its values or uniqueness, I want to remind them that if you can just prompt-create a product, so can the AI provider. In fact, they might have better resources to create a competing product if it displays any sign of success (see Figma). While it looks like plenty of people are benefitting from AI today, all this information is being shared with AI providers. We are giving them full access to our thought process. When you include them in your workflow, you are basically providing them with a step by step approach on how to do your job. Don’t be surprised when you see a native Antropic/OpenAi project management application suite. Or a CRM, or any software that is trying to integrate with AI and may experience success. A few years back, when I worked in Customer Service Automation, we discovered that most companies used Zendesk to manage their customer service. Since customers mainly contacted support via email, an intentional database had been built that tracked users through their shopping experience throughout the web. While so much could be done with that data, like identifying “problematic” customers, or recommending products based on their history, we ended up finding something more helpful. We could easily detect a pattern of issues for certain shipping carriers. We could see when UPS was having delays in certain cities, or when Fedex was having technical issues when updating the last mile status. None of these things were features designed or provided by anyone. However, having access to businesses’ data gave us insight where we had none before. That became a feature for us, only because we were not competitors to all these online retailers. When you expose your company's internal data to a potential competitor, don’t be surprised when they build a competing business to rival you.

business Business data Data
0 views