#UserHostile hashtag - Bluesky

18 hours ago

Phase 3: Profit — The GitHub Story - ByteHaven - Where I ramble about bytes Part of the ongoing Big Tech’s War on Users series. Earlier this past week, GitHub announced that starting April 24th, your Copilot interaction data will be...

Four fuses. One week. All burning.
OpenAI raised $122B and still won't turn a profit until 2030. Agents are deleting inboxes and want root access to your machine. Your RAM is three times what it was because hyperscalers bought the supply chain out from under you. And nobody in the industry can […]

1 4 0 0

ppb1701.bsky.social

@ppb1701.bsky.social

3 days ago

Phase 1: Acquire developer trust. Phase 2: ???. Phase 3: Your code trains our AI by default. The GitHub story.

blog.ppb1701.com/phase-3-prof...
#bigtech #blog #github #microsoft #privacy #selfhosting #userhostile #opensource #ai

2 0 0 0

Phase 3: Profit — The GitHub Story _Part of the ongoing_ _Big Tech’s War on Users_ _series._ Earlier this past week, GitHub announced that starting April 24th, your Copilot interaction data will be used to train Microsoft’s AI models. By default. Unless you find the setting and turn it off. It’s worth pausing on that word — _default_. Because in 2018, when Microsoft acquired GitHub for $7.5 billion, developers were nervous. This was Microsoft — the company that called Linux “a cancer,” bundled Internet Explorer to kill competition, and had spent decades treating developers as a captive audience. Handing them the world’s largest code host felt like handing the keys to the commons over to a landlord. Satya Nadella knew exactly what he was walking into. So he made a promise. Specific, public, on the record — from the official Microsoft blog: > “We are committed to being stewards of the GitHub community, which will retain its developer-first ethos, operate independently and remain an open platform.” And then he invited accountability — his exact words: > “When it comes to our commitment to open source, judge us by the actions we have taken in the recent past, our actions today, and in the future.” Okay. Let’s do that. ## Phase 1: Collect Underpants (2018) The acquisition announcement was a carefully managed reassurance campaign. Nat Friedman — a respected open source figure, founder of Xamarin — was installed as CEO specifically to signal continuity. The message was consistent across every channel: GitHub would remain independent. Developer-first. A neutral commons. Not a Microsoft product. It worked. Developers who had started migrating to GitLab came back or stayed. The GitHub community largely gave Microsoft the benefit of the doubt. Nadella had spent years rehabilitating Microsoft’s relationship with open source and this felt like the logical extension of that goodwill. The trust was real. It was also load-bearing. ## Phase 2: ??? (2025) Seven years later, in August 2025, GitHub CEO Thomas Dohmke announced he was stepping down. Microsoft did not announce a replacement. Instead, GitHub was folded into Microsoft’s CoreAI division under Jay Parikh, executive vice president. No fanfare. No “we’re evolving the structure.” Just a reorganization that ended GitHub’s operational independence and positioned it explicitly as a component of Microsoft’s AI infrastructure. The independence promise didn’t expire with an announcement. It just quietly stopped being true. ## Phase 3: Profit (2026) On March 25th, GitHub announcedthat starting April 24th, interaction data from Copilot Free, Pro, and Pro+ users will be used to train and improve Microsoft’s AI models. By default. Unless you find the setting and turn it off. What counts as interaction data is worth reading carefully: * Inputs sent to GitHub Copilot * Outputs accepted or modified by you * Code context surrounding your cursor position * Comments and documentation you write * File names, repository structure, and navigation patterns * Interactions with Copilot features * Feedback — thumbs up/down ratings And who gets it? GitHub, and its affiliates. Which, as they specifically note in their updated privacy statement, includes Microsoft. If you’re on Copilot Business or Enterprise, none of this applies to you. Your corporate contract prohibits it. You’re protected. If you’re an individual developer — open source contributor, indie hacker, someone who touches their employer’s codebase from a personal account — you’re in until you opt out. The responsibility to do so is entirely yours. ## The Reversal Is Documented This isn’t a new policy being introduced. It’s an old promise being broken. In November 2025 — four months ago — a GitHub maintainer answered this exact question in the community forums: > “By default, GitHub, its affiliates, and third parties do not use your data (prompts, suggestions, code snippets) for AI model training. This setting cannot be enabled — it’s off and stays off.” Cannot be enabled. Off and stays off. That answer is now wrong. As of April 24th, for 26 million users, it’s on and stays on unless you turn it off yourself. ## The Windows Connection Here’s where the timing becomes something more than coincidence. Five days before the GitHub announcement, on March 20th, Windows president Pavan Davuluri published what I called a landmark “we hear you” post — promising to fix Windows 11. Less Copilot. Fewer ads. Movable taskbar. A calmer OS. Same week. Different audience. Different message. The developer and enthusiast community responded cautiously but positively. Microsoft was listening. Things were changing. Same week. Same company. One hand pulling Copilot back from the Start menu. The other hand quietly routing your code into the Copilot training pipeline. I wrote in that post that a calmer cage is still a cage. I didn’t know quite how literal that was going to get, quite that fast. The consumer audience gets the soothing blog post. The developer audience gets the policy update. Different pressure release valves for different communities — but the same direction of travel underneath. ## The Class System Nobody Is Talking About Enough GitHub’s own FAQ explains the individual/enterprise split plainly: business and enterprise customers have contracts that prohibit this use of their data. Those agreements are honored. Individual users — the open source contributors, the hobbyists, the people who _built the community that made GitHub worth $7.5 billion_ — get opted in by default. One commenter on Hacker News framed it generously: this is standard B2B SaaS practice. Enterprise DPAs include ML training carve-outs that free users don’t get. That’s true. It’s also exactly the point. The community that was promised stewardship and independence is getting consumer-tier treatment. The enterprise customers paying the big contracts get the protection that stewardship actually looks like. The people who made this platform worth buying are not the people being protected by it. ## What The Developer Community Is Doing Migration guides to Codeberg and Forgejo started appearing within hours of the announcement. Not rage-tweets. Documentation. Step-by-step instructions for moving repositories, redirecting links, rebuilding CI pipelines. This matters because the switching cost is genuinely high. These aren’t casual users threatening to delete an app. These are people with years of commit history, issue threads, CI configurations, and public presence tied to GitHub. When developers with that much to lose start writing exit documentation rather than filing angry issues, that’s a meaningful signal. Worth noting: a significant chunk of the GitHub open source community is already on Linux or Mac. The Windows “we hear you” post wasn’t aimed at them and didn’t land for them. The Microsoft goodwill tour doesn’t apply to the audience most affected by this policy change. They were already skeptical. This confirms it. ## The Private Repository Sleight of Hand GitHub draws a careful distinction between data “at rest” and active interactions. Your private repository contents, sitting on their servers, are not used for training. But if you’re actively working in that private repository with Copilot enabled — your code context, your cursor position, your comments, your file structure — that’s interaction data. That’s in scope. Unless you opt out. If you’re an individual developer working on a personal account who sometimes touches your employer’s proprietary codebase, that’s worth sitting with for a moment. ## The Opt-Out Friction Problem Community reaction on Reddit and Hacker News has been pointed, with a consistent complaint beyond the policy itself: the opt-out instructions are confusing and inconsistent across different GitHub documentation pages. The setting is at: GitHub Account Settings → Copilot → Privacy → “Allow GitHub to use my data for AI model training” → Disabled. If you use multiple GitHub accounts, you need to do it for each one separately. Data retention if you don’t opt out: up to five years, though often shorter per GitHub’s FAQ. ## Nadella Said Judge Them By Their Actions Eight years of receipts: * 2018: GitHub will operate independently, retain developer-first ethos, remain open platform * 2021: Nat Friedman steps down, Thomas Dohmke takes over — independence maintained on paper * 2025: Dohmke out, GitHub absorbed into CoreAI, independence ends without announcement * 2026: Individual developer code routes into Microsoft AI training by default The promise was load-bearing. The trust it bought was real, and it’s been spent. Twenty-six million developers are on a platform whose terms just materially changed. Nadella said judge them by their actions. He was right to invite that. It’s the only honest standard. ## What To Do **Opt out now** if you haven’t: GitHub Settings → Copilot → Privacy → disable “Allow GitHub to use my data for AI model training.” Do it for every account. **Consider your architecture.** GitHub as a primary with no alternative is a single point of trust failure that just demonstrated it can fail. GitHub as a mirror to a self-hosted Gitea or Forgejo instance, or to Codeberg for public repos, is a different relationship entirely — you’re using their infrastructure as a utility without depending on their goodwill. For what it’s worth, my setup is self-hosted Gitea as the actual source of truth — my NixOS configs and other bits live there, with GitHub as a mirror mostly for shareability and discoverability. I’d already flipped the opt-out before writing this. But I’m now actively looking for a different public-facing home, and rethinking how I handle private repos entirely. I’ll cover what I actually decided to do in an upcoming home server post — but the short version is that this policy change is the thing that finally pushed me to treat private code like any other sensitive data on my infrastructure rather than something I hand to a third party and hope for the best. If you want to go the self-hosted route, Part 11 of my home server series covers setting up Gitea. If Codeberg is more your speed — hosted, nonprofit, EU-based, no VC money — there’s a solid practical migration guide on DEV written this week by someone who just did it. And if you want to self-host Forgejo specifically, this walkthrough covers the full process including CI setup via Forgejo Actions. The official Codeberg migration docs are also worth bookmarking — built-in GitHub import handles issues, PRs, labels and releases with checkboxes, not command line gymnastics. **Watch the policy.** It changed once with 30 days notice. The precedent is set. The commons was never really free. It was subsidized by developer trust that accumulated over decades and has now been drawn down in service of a $7.5 billion acquisition’s AI strategy. Judge them by their actions. Nadella’s words. Still good advice. _Find me on Mastodon at_ _@ppb1701@ppb.social_ _Part of the ongoing_ _Big Tech’s War on Users_ _series._

3 days ago

Phase 1: Acquire developer trust. Phase 2: ???. Phase 3: Your code trains our AI by default. The GitHub story.

blog.ppb1701.com/phase-3-profit-the-githu...

#bigtech #blog #github #microsoft #privacy #selfhosting #userhostile #opensource #ai

2 4 0 0

Netflix Got $2.8 Billion Last Month. Now It Wants More of Yours. Last month, I closed out the Warner Bros. bidding war saga with this line: > _"Whether it's Netflix or Paramount holding the keys to Warner Bros., that part doesn't change. We're still paying $18/month for a streaming service that drops three episodes of a show and then makes us wait six months for the rest. The consolidation is happening, the prices will keep going up, and the content will keep getting spread thinner."_ > I didn't think it would take less than a month to be proven right. On Thursday, March 26, Netflix quietly updated their help page with new pricing across every single plan. No press release. No announcement. Just: here are your new prices, effective today for new subscribers, rolling out to existing members over the coming weeks. ## What You're Paying Now Here's the damage: Plan Old Price New Price Change Standard with Ads $7.99/mo $8.99/mo +$1 Standard (no ads) $17.99/mo $19.99/mo +$2 Premium (4K, 4 streams) $24.99/mo $26.99/mo +$2 Extra Member (ad-supported) $5.99/mo $6.99/mo +$1 Extra Member (ad-free) $8.99/mo $9.99/mo +$1 This is the **second price hike in just over a year.** The last one hit in January 2025. If you want Netflix without ads, the cheapest you can get it is now **$19.99 a month.** If you want 4K, you're looking at **nearly $30.** ## Netflix's Statement Is a Work of Art in Corporate Non-Language Netflix's statement to Variety: _"As we deliver more value to our members, we are updating our prices to enable us to reinvest in quality entertainment and improve their experience by updating our prices."_ They said "updating our prices" **twice** in the same sentence. That's not an accident. That's what happens when your PR team has to explain a price increase without using the word "increase." The "more value" argument is doing a lot of heavy lifting here. Yes, Netflix has added video podcasts, live events, and sports content. They've also been spending $20 billion on content in 2026 alone — up from $18 billion in 2025. But nobody asked for video podcasts. And the core experience — dropping partial seasons, algorithmic recommendations that feel like they're actively avoiding what you want, the endless scroll — hasn't materially improved. ## Let's Talk About the $2.8 Billion Here's the part I can't get past. Netflix just walked away from the Warner Bros. acquisition last month and **collected a $2.8 billion termination fee** in the process. Netflix CFO Spence Neumann stood at an investor conference and said — and I want you to really hear this — _"Now we move forward, and we move forward with $2.8 billion in our pocket that we didn't have a few weeks ago."_ Cool. Great. Terrific. Now they've turned right around and raised prices on every subscriber in America. I'm not suggesting there's a direct causal link — the price hike was telegraphed in Netflix's Q4 earnings guidance in January, before the deal collapsed. But the optics of walking away from an $82.7 billion acquisition with a $2.8 billion consolation prize and then immediately coming for your customers' wallets next billing cycle... that's a specific kind of audacity. ## The Ad-Supported "Escape Hatch" Is Getting More Expensive Too This is the part that should actually annoy you. When Netflix launched its ad-supported tier, the pitch was pretty clear: if you can't stomach the price increases on the premium tiers, here's a cheaper option. Accept some ads, pay less. Fine. That tier just went from $7.99 to **$8.99.** A $1 increase on a plan that still shows you the same ads — research from Parks Associates found that seven in ten viewers say the same ads repeat too often on streaming services. You're paying more for the same annoying experience. (And if you've been following the broader war on users who just want to watch something without being monetized to death, this pattern should feel very familiar.) And the gap between ad-supported and ad-free is now **$11 a month.** If you don't want ads, that's $11 extra every single month. The "cheap" tier isn't a path down anymore — it's just a slower path to the same destination. ## You're Also Paying More for a Worse Viewing Experience The price is only part of the story. The _product_ has also gotten worse in ways that never make the press release. **The binge model is quietly dying.** Netflix built its entire identity on dropping full seasons at once. That was the deal — you pay, you watch, you don't have to wait. Now they're increasingly releasing in batches or on weekly schedules. You get half a season, then wait a month for the back half. Or you get one episode a week like it's 1987 and you have rabbit ears. They didn't announce this as a policy change. They just started doing it. **The gap between seasons is getting longer.** Stranger Things is the most egregious example — fans waited over three years between Season 4 (July 2022) and the final season (November 2025). A show that's been running since 2016, and it took nearly a decade to produce five short seasons. Now, plenty of people have figured out the workaround: cancel, wait for the drop, resubscribe, binge, cancel again. Netflix knows you're doing this. They've _always_ known. And their response hasn't been to shorten the gaps or reward loyalty — it's been to make the content calendar unpredictable enough that you're never quite sure when it's safe to cancel without missing something. The churn game is a feature, not a bug. **And if a show survives long enough to get traction, there's a decent chance it gets cancelled at Season 3.** Not because it failed — sometimes right when it's hitting its stride. Netflix has developed a well-documented pattern of greenlighting shows, letting them build an audience over two or three seasons, and then axing them right around the point where they'd have to start paying the cast and crew significantly more. The show doesn't get an ending. The audience doesn't get closure. _Santa Clarita Diet._ Three seasons. A genuinely weird, funny, beloved show — Season 3 scored 100% on Rotten Tomatoes — with a cliffhanger ending that will never be resolved. Drew Barrymore and Timothy Olyphant spent three seasons building something people actually cared about, and Netflix pulled the plug right before the story paid off. Creator Victor Fresco found out his show was cancelled while sitting in the edit suite — an assistant producer walked in to tell him the sets were being dismantled. That's how he heard. Fresco had deliberately ended Season 3 on a cliffhanger to try to force Netflix's hand — _"We didn't want to make it easier for them to cancel us,"_ he said. It didn't matter. They booked the subscriber acquisition value from the marketing push, counted the streams, and sent someone to tear down the sets. The fans are still out there, years later, knowing they'll never find out what happened. That's not a content strategy. That's using your audience as a resource and discarding them when the math changes. ## And Then There's the Chef's Kiss That Ties the Whole Thing Together In 2017, Netflix posted a tweet that became something of a brand identity statement: _"Love is sharing a password."_ They meant it. CEO Reed Hastings said the same year that password sharing was _"a positive thing, not a negative thing."_ For years, sharing your Netflix login with a kid at college, a parent across town, or a new partner was practically a cultural ritual. Netflix knew. Netflix encouraged it. Netflix built an audience of over 100 million password-sharing households on the back of that goodwill. Then in 2023, they killed it. Suddenly sharing a password was a terms-of-service violation, and adding someone outside your household would cost you extra — on top of the subscription you were already paying. The tweet is still out there. The goodwill it was built on is not. So to recap: they got you hooked on a generous, flexible product. Raised prices. Cracked down on the sharing they told you was love. Switched from full season drops to batches and weekly episodes. Cancelled the shows you cared about before they could finish their stories. Made the content calendar unpredictable enough that you can't even strategically cancel anymore. And now they want two more dollars a month. ## The Pattern Is the Point Netflix isn't an outlier. Every major streamer has raised prices in the past two years. But Netflix's pace is worth documenting. Netflix's Standard plan history, roughly: * **2019:** ~$12.99/month * **2022:** $15.49/month * **Early 2025:** $17.99/month * **Today:** $19.99/month The mid-tier plan has gone up roughly **54% in five years.** Meanwhile, Netflix's Q4 2024 earnings set subscriber records. The company is projecting $50+ billion in revenue for 2026. They're not raising prices because they're struggling. They're raising prices because they can. ## What You Can Actually Do Honestly? Not a lot. Netflix has the content. They know they have the content. Your leverage is cancellation, and they also know that most people who cancel come back. But if you're paying for Premium and only using two screens — drop to Standard. If you haven't audited your streaming subscriptions lately, now is a good time. The average household holds nearly six streaming subscriptions; that's a lot of $2 increases adding up across a lot of services. Or don't. They've already factored your inertia into their revenue projections. That's the most honest thing I can tell you. _Existing subscribers will get email notice one month before the new prices hit their billing cycle. New subscribers see the new prices starting today._ **Got thoughts on this mess? Find me on Mastodon at****@ppb1701@ppb.social**

5 days ago

Netflix Got $2.8 Billion Last Month. Now It Wants More of Yours.

blog.ppb1701.com/netflix-got-2-8-billion-...

#blog #netflix #streaming #streamingwars #bigtech #userhostile #enshitification

0 2 1 0

ppb1701.bsky.social

@ppb1701.bsky.social

6 days ago

Bitwarden Doubled Their Price. I'd Already Left. Here's What You Missed.

blog.ppb1701.com/bitwarden-do...

#bitwarden #passwordmanager #selfhosting #userhostile #blog

0 1 0 0

Bitwarden Doubled Their Price. I'd Already Left. Here's What You Missed. Back in January, I was two days into setting up Vaultwarden when Bitwarden sent me their annual Data Privacy Week survey. What browser do you use? What email service? What VPN? What are your concerns about AI and your data? I filled it out. Checked every box about keeping AI away from my personal data. Sent it in. Went back to configuring my self-hosted instance. The irony wasn't lost on me — a password manager asking how I feel about AI and data privacy, while I was in the middle of moving my passwords off their servers entirely. Six days later, on January 21st, they doubled their Premium price. I found out about it recently — because someone sent me a Fast Company article from February 1st. Which itself took eleven days after the announcement for anyone to even notice it was worth writing about. That tells you everything about how quietly this was done. ## What Actually Happened Bitwarden's Premium plan went from $10/year to $19.80/year — their first price increase in ten years. The announcement was a blog post about new features — vault health alerts, password coaching, more attachment storage — with the price change mentioned almost as an afterthought. No dedicated email. No direct "hey, your price is doubling" communication to paying customers. I follow them on Mastodon and checked their Facebook — (yes, I know, Facebook, but when your local government refuses to use anything else for announcements, you end up keeping the account) — if they announced it on social, it wasn't given any prominence. Either it's not there or it's buried so deep it doesn't count as communication. Existing customers will find out 15 days before their renewal. Not in an email that says "your price is going from $10 to $19.80." In an email that says "$1.65/month, billed annually." Here's the thing — Bitwarden has never offered monthly billing. Not once in ten years. It has always been annual only. So if you're a paying customer who isn't paying close attention, that email looks less like a price hike and more like... taxes? A currency adjustment? Nothing to see here, auto-renew. That's not an accident. That's the design. ## My Situation Specifically I paid Bitwarden on the way out. Deliberately. I'd been a satisfied customer, I was leaving for my own reasons, and I wanted to say thank you — a coffee or two's worth of goodwill. I left on good terms. In return, as a recent paying customer, I received zero communication about the price doubling. I would have found out at renewal — nearly a year after the fact — via a deliberately confusing email that never actually states the annual price. I'd have been sitting there thinking "did they add taxes?" and probably just let it roll. I'm not angry about $20 a year. That's not crazy money for a password manager. But I am calling out the way this was done, because this is exactly the pattern I document in this series. ## The Pattern Bitwarden isn't Google or Apple. They're not a monopoly. But the playbook rhymes: * **Bury the bad news** — hide a price doubling inside a feature announcement * **Obscure the actual number** — list a monthly price for an annual-only product * **Minimize direct communication** — let customers find out at renewal, not upfront * **Bank on friction** — password manager migration is painful, most people won't bother That last one is the real bet. They know switching costs are high. They know most customers will grumble and auto-renew. And the communication strategy is specifically engineered to minimize the number of people who even grumble. I went from "on the fence about whether this was worth writing about" to "this is getting shady" in about ten minutes of looking at the details. When a company's pricing communication is specifically designed so customers don't notice what's happening — that's not a mistake. That's a choice. ## Why I'm Not Scrambling Here's what's different about my situation: I don't have to decide anything. The Vaultwarden instance has been running since mid-January. Browser extension, mobile app, desktop client. Zero issues. My passwords live on my server, not theirs. I'll be sunsetting my Bitwarden cloud presence soon. Not in anger — just tidying up. The cloud account was always going to be a temporary failsafe while I validated the setup. It's validated. Time to close the door properly. When the price hike landed, I wasn't scrambling. I was just watching, mildly interested, from somewhere they can no longer reach. That's the actual value of self-hosting. Not the features. Not the money saved. The fact that when a company does something you don't like, you're not stuck. ## What You Can Do Most people aren't going to self-host a password manager. That's fine. But if you're a Bitwarden Premium subscriber who just found out about this — especially _how_ it was communicated — you have options. Proton Pass and 1Password are both worth a look, and at their current pricing they're not dramatically more expensive than the new Bitwarden rate. And if you want to actually own your password data: here's how I did it. **Did you know about the Bitwarden price increase? Did you get a direct email about it? Find me on Mastodon at@ppb1701@ppb.social — because I didn't, and I'm genuinely curious how many paying customers are still in the dark.**

6 days ago

Bitwarden Doubled Their Price. I'd Already Left. Here's What You Missed.

blog.ppb1701.com/bitwarden-doubled-their-...

#bitwarden #passwordmanager #selfhosting #userhostile #blog

0 3 0 0

1 week ago

When I started this series I figured I'd be documenting slow burns. The gradual squeeze. The quietly updated terms of service. The feature that disappears in a patch note nobody reads.
Nobody told Nvidia.
Hello Phobos.

https://blog.ppb1701.com/the-crowbar-on-pandoras-box

#ai #dlss #gaming […]

0 5 0 0

1 week ago

New post in the Big Tech's War on Users series. Microsoft just promised to make Windows 11 better. Two days later they quietly reached into your machine and turned off a performance fix the community had found themselves. Oh and the official fix requires a laptop-only Intel chip that launched in […]

0 5 0 0

Jensen's Apology Tour Changes Exactly Nothing It's been a couple of days and I thought perhaps we'd move on — next entry in Big Tech versus users, something for my server, maybe even touch on Microsoft promising to make Windows 11 better (which I'm still liable to circle back to, because that is a whole thing in and of itself). Instead, this happened today. So Jensen Huang went on Lex Fridman's podcast. "I don't love AI slop myself," he said. Very relatable. Very human. The CEO of a $4 trillion company, empathizing with the people his company told were "completely wrong" just days earlier. Here's the thing though: you don't get to do the AI slop sympathy tour when your company is simultaneously lobbying for special permissions to sell AI chips to China and signing every hyperscaler contract that moves. The balance sheet is the honest statement. The podcast is PR. I wrote last week about the dual-5090 demo rig — a system pushing $10,000-12,000 — being the tell. Nvidia chose to unveil a "gaming enhancement" feature on hardware that almost no gamer will ever own, at an AI infrastructure conference, in front of investors and enterprise customers. That wasn't an accident. That was the audience. And here's what the walk-back still doesn't address: **Nvidia's bread and butter isn't you.** Gaming revenue is a rounding error next to datacenter. The hyperscalers — Amazon, Microsoft, Google — account for the overwhelming majority of what Nvidia actually makes. Every H100 and GB200 that ships to an AI data center is a direct contributor to the supply crunch and price inflation that's put a 5090 at $4,232 on Amazon right now — more than double its $1,999 MSRP — and analysts think $5,000 is coming by late 2026. Jensen saying he doesn't love AI slop is a bit like an oil executive saying he doesn't love high gas prices. He's technically allowed to feel that way. Doesn't change the quarterly earnings call. What really stings is what the demo _could_ have been. Wheel out a Legion Go or an ROG Ally. Something a person can actually order online today. Show DLSS doing for a handheld what it genuinely does best — making constrained hardware punch way above its weight class. That's a crowd that would have been too busy checking preorder availability to complain about Grace Ashcroft's makeup. Instead they brought a server to a gaming party and then acted surprised when gamers weren't grateful. To make it worse, it turned out Nvidia's own GeForce evangelist effectively contradicted Jensen's technical claims mid-damage-control cycle — the "it's not post-processing, it's geometry-level generative control" defense crumbled when his own team's description of how the tech actually works told a very different story. The Lex Fridman charm offensive will probably work on the news cycle. The discourse moves on, the memes age out, and by fall when DLSS 5 is supposedly optimized for a single GPU most people will have forgotten the demo rig and the "completely wrong" quote. But the underlying dynamic doesn't move. Nvidia needs the AI bubble to keep inflating to justify $1 trillion in demand projections. That bubble is funded by the same companies consuming GPU supply and driving component prices up. Consumers are downstream of all of it — paying more, finding less, and being told to wait for the optimized version. Jensen's empathy is free. The 5090 is still $4,232. **Is the Lex Fridman walk-back enough to change your mind, or does the balance sheet tell the real story? Find me on Mastodon at@ppb1701@ppb.social and let's talk about it.**

1 week ago

Jensen said "I don't love AI slop" on Lex Fridman this week.
The 5090 is still $4,232.
New post on why the apology tour changes exactly nothing

blog.ppb1701.com/jensens-apology-tour-cha...

#Nvidia #DLSS5 #Gaming #AI #GPU #blog #userhostile

0 5 0 0

1 week ago

Jensen announced everything Monday. We're just now finding out how deep it goes.
New post: "You Load Sixteen Tokens, What Do You Get?"
The mine got safer. The model stayed the same.

blog.ppb1701.com/you-load-sixteen-tokens-...

#ai #bigtech #blog #compensation #tokens […]

0 3 0 0

2 weeks ago