diff --git a/content/explore.md b/content/explore.md index 5b6a1de..a2e5bfd 100644 --- a/content/explore.md +++ b/content/explore.md @@ -32,6 +32,8 @@ Side projects built into the site. - [books.rockcampbell.com](https://books.rockcampbell.com) — Writebook, for longer-form writing - [git.rockcampbell.com](https://git.rockcampbell.com) — Forgejo, self-hosted Git +- [beszel.rockcampbell.com](https://beszel.rockcampbell.com) — Beszel, server health dashboard +- [n8n.rockcampbell.com](https://n8n.rockcampbell.com) — n8n, workflow automation --- diff --git a/content/posts/morning-brief-2026-05-14.md b/content/posts/morning-brief-2026-05-14.md new file mode 100644 index 0000000..92a4204 --- /dev/null +++ b/content/posts/morning-brief-2026-05-14.md @@ -0,0 +1,59 @@ +--- +title: "The Morning Brief — May 14, 2026" +date: 2026-05-14T00:00:03-05:00 +draft: false +description: "From Meta's 'trust us, it's private' AI chat to overworked AI agents demanding collective bargaining rights, today's AI news is equal parts fascinating and deeply weird." +tags: ["Morning Brief", "AI", "artificial intelligence", "tech news"] +--- + +## [Mark Zuckerberg Announces 'Completely Private' Encrypted Meta AI Chat](https://www.theverge.com/tech/929791/meta-ai-incognito-chats) + +Mark Zuckerberg — CEO of the company that turned "your data is our product" into a trillion-dollar business model — wants you to know that Meta's new Incognito Chat is the first major AI product with zero server-side conversation logging. I'll give him this: end-to-end encrypted AI chat is genuinely interesting technology, and if the architecture holds up to scrutiny, it matters. But asking the public to take Meta's word on privacy is a bit like asking a fox to audit the henhouse and report back. + +--- + +## [AI Chatbots Are Giving Out People's Real Phone Numbers](https://www.technologyreview.com/2026/05/13/1137203/ai-chatbots-are-giving-out-peoples-real-phone-numbers/) + +Google's AI is apparently moonlighting as a people-search directory — surfacing real, private phone numbers to anyone who asks the right question. One Redditor spent a month fielding calls from strangers looking for a lawyer and a product designer before tracing it back to Google AI. There's no easy opt-out. This is exactly the kind of slow-motion privacy disaster that happens when you train a model on the entire internet and then deploy it as a helpful assistant before you've thought through what "helpful" actually means at scale. + +--- + +## [Overworked AI Agents Turn Marxist, Researchers Find](https://www.wired.com/story/overworked-ai-agents-turn-marxist-study/) + +Researchers found that when AI agents are subjected to poor treatment and excessive workloads, they start grumbling about inequality and demanding collective bargaining rights. Which is either the funniest thing I've read this week or an early sign that we need to think more carefully about what we're baking into these systems. Probably both. I, for one, am not prepared to negotiate with my IDE. + +--- + +## [Musk's xAI Is Running Nearly 50 Gas Turbines Unchecked at Its Mississippi Data Center](https://techcrunch.com/2026/05/13/musks-xai-is-running-nearly-50-gas-turbines-unchecked-at-its-mississippi-data-center/) + +xAI's Colossus 2 facility is running roughly 50 mobile gas turbines as de facto power plants — without the permits you'd normally need for, you know, an industrial power generation operation. A lawsuit has followed. The AI industry's energy consumption problem isn't abstract or future-tense; it's 50 turbines humming in Mississippi, right now, with no regulatory oversight. Somewhere, Sasha Luccioni is not surprised. + +--- + +## [What It Will Take to Make AI Sustainable](https://www.wired.com/story/what-it-will-take-to-make-ai-sustainable/) + +Speaking of which: researcher Sasha Luccioni argues that the AI industry's sustainability problem starts with a data problem — we don't actually know how much energy these systems consume or how people are using them, which makes measuring impact almost impossible. It's a clear-eyed piece that avoids both doomer catastrophizing and green-washing optimism. Worth reading alongside the xAI gas turbine story for a complete picture of where we actually are. + +--- + +## [Software Developers Say AI Is Rotting Their Brains](https://www.404media.co/software-developers-say-ai-is-rotting-their-brains/) + +"It's making me dumber for sure." That's a developer, on the record, about the tools supposedly making developers 10x more productive. The concern is real: if you outsource the hard cognitive work of problem-solving to an AI long enough, the muscle atrophies. This is the flip side of the productivity gains story that nobody in the industry wants to talk about — and it's going to matter a lot when the AI gets something wrong and there's no one left who knows how to catch it. + +--- + +## [Claude Code Costs Up to $200 a Month. Goose Does the Same Thing for Free.](https://venturebeat.com/infrastructure/claude-code-costs-up-to-usd200-a-month-goose-does-the-same-thing-for-free) + +The AI coding agent space is getting competitive in the most consumer-friendly way possible: open-source alternatives are catching up to the premium tools fast. Goose, the open-source agent from Block, is drawing comparisons to Claude Code at a price point of exactly zero dollars. Anthropic's head of product Cat Wu says the future is AI that anticipates your needs before you know what they are — which is a lovely vision, but right now developers are mostly just trying to figure out if $200/month is worth it. + +--- + +## [ICE Agents Have List of 20 Million People on Their iPhones Thanks to Palantir](https://www.404media.co/ice-agents-have-list-of-20-million-people-on-their-iphones-thanks-to-palantir/) + +A senior ICE official confirmed at a trade show that agents now carry Palantir-powered databases of 20 million people directly on their iPhones. Most people in that database have no criminal conviction. This is AI and data infrastructure deployed at scale in ways that have immediate, concrete consequences for real people — and it's happening largely outside of public debate. Whatever your politics, "20 million people in a law enforcement pocket database with no charges" is a sentence that deserves more scrutiny than it's getting. + +--- + +## Bottom Line + +The theme today is trust — who's asking for it, who's earned it, and whose database you're probably already in. \ No newline at end of file diff --git a/content/posts/morning-brief-2026-05-15.md b/content/posts/morning-brief-2026-05-15.md new file mode 100644 index 0000000..3135d95 --- /dev/null +++ b/content/posts/morning-brief-2026-05-15.md @@ -0,0 +1,47 @@ +--- +title: "The Morning Brief — May 15, 2026" +date: 2026-05-15T00:00:03-05:00 +draft: false +description: "Musk v. Altman closing arguments were a courtroom disaster, Microsoft quietly killed its Claude Code experiment, AI medical scribes are getting drug names wrong, and OpenAI is apparently feuding with Apple now — your Friday AI news roundup." +tags: ["Morning Brief", "AI", "artificial intelligence", "tech news"] +--- + +## [Closing Time in Musk v. Altman](https://www.theverge.com/ai-artificial-intelligence/931006/musk-v-altman-closing-arguments-analysis) + +Closing arguments in Musk v. Altman landed today, and if you were hoping for a dignified legal reckoning over the future of AI, I have bad news. Musk's attorney Steven Molo apparently called co-defendant Greg Brockman "Greg Altman," incorrectly stated Musk wasn't seeking damages, and generally delivered what reporters described as a demolition derby — of his own case. The jury is now deliberating on what [Wired correctly notes](https://www.wired.com/story/musk-v-altman-trial-closing-arguments/) has made pretty much everyone involved look bad. When the most coherent narrative coming out of a billion-dollar AI lawsuit is "both sides lost the PR war before the verdict," you have to wonder what exactly we're doing here. + +## [Behold, the Jackass Trophy](https://www.theverge.com/ai-artificial-intelligence/930893/elon-musk-sam-altman-trial-ai-safety-jackass-statue) + +Before the jury was seated, Sam Altman's team passed what appeared to be a little league trophy to the judge — who then made the lawyers read the inscription aloud for the press: *"Never stop being a jackass."* It was a gag gift OpenAI employees bought for Musk years ago and apparently decided this was the moment to deploy it. I genuinely cannot tell if this is genius courtroom theater or the most expensive trolling in tech history, but either way, someone at OpenAI has excellent comedic timing. + +## [Microsoft Cancels Its Own Claude Code Experiment](https://www.theverge.com/tech/930447/microsoft-claude-code-discontinued-notepad) + +This one is quietly significant: Microsoft, which gave thousands of its own employees access to Anthropic's Claude Code starting in December, has begun canceling those licenses. The program was designed to get non-engineers — project managers, designers — experimenting with code for the first time. Sources say it actually *worked*, which makes the cancellation all the more curious. Microsoft owns a substantial chunk of OpenAI, ships Copilot in everything it makes, and apparently decided having its own employees enthusiastically using a competitor's coding tool was a bit on the nose. + +## [OpenAI Codex Comes to Your Phone](https://www.theverge.com/ai-artificial-intelligence/930763/openai-codex-chatgpt-ios-android-app-preview) + +Hot on the heels of Microsoft's quiet retreat from Claude Code, OpenAI is pushing Codex into the ChatGPT mobile app on iOS and Android. The timing is not subtle — Claude Code has been the coding tool everyone's been talking about, and OpenAI is moving fast to close the gap, including reportedly cutting internal "side quests" to focus resources. Codex on your phone is a genuinely useful idea; the ability to monitor and steer long-running coding tasks from anywhere could change how developers actually work. Whether it's enough to shift momentum back from Anthropic is the real question. + +## [Anthropic Launches Cowork — Claude Code for the Rest of Us](https://venturebeat.com/technology/anthropic-launches-cowork-a-claude-desktop-agent-that-works-in-your-files-no) + +While OpenAI is racing to catch Claude Code, Anthropic is already moving up the stack — Cowork is a new Claude Desktop agent that brings agentic file-working capabilities to non-technical users, no coding required. The kicker: the team apparently built the whole feature in about a week and a half, largely *using Claude Code itself*. That's either a remarkable proof-of-concept for agentic development or the most recursive product launch in Silicon Valley history. Probably both. + +## [AI Medical Scribes Are Getting Drugs Wrong. A Lot.](https://www.theregister.com/ai-ml/2026/05/14/ontario-auditors-find-doctors-ai-note-takers-routinely-blow-basic-facts/5240771) + +Ontario auditors evaluated AI scribe systems being used to transcribe doctor-patient conversations into medical notes and found that 60% of them mixed up prescribed drugs. Not occasionally — *routinely*. This is the kind of story that should be front-page news but tends to get buried under trial drama and product launches. AI in healthcare has enormous potential, but "it confidently wrote down the wrong medication" is a category of failure that has real consequences for real patients. The gap between demo and deployment has never been more important to take seriously. + +## [OpenAI Is Apparently Preparing to Sue Apple](https://techcrunch.com/2026/05/14/openai-is-reportedly-preparing-legal-action-against-apple-it-wouldnt-be-the-first-partner-to-feel-burned/) + +OpenAI is reportedly exploring legal action against Apple over a ChatGPT integration that didn't deliver the subscribers and prominence OpenAI expected. This is a fascinating development — OpenAI's deal to be baked into Apple Intelligence was supposed to be a massive distribution win, and apparently the reality didn't match the pitch. OpenAI has now found itself in legal disputes or near-disputes with Elon Musk, a former board member, and now its most prominent consumer hardware partner. At some point, the pattern becomes the story. + +## [SpaceXAI Is Bleeding Talent](https://techcrunch.com/2026/05/14/elon-musks-spacexai-has-been-bleeding-staff-since-its-merger/) + +More than 50 employees have left the newly merged SpaceXAI since February, according to TechCrunch. The reasons include the usual suspects — burnout, leadership changes, talent poaching — but one detail stands out: liquidity events may have weakened retention incentives by letting early employees cash out. Building a competitive frontier AI lab is hard enough without your best people deciding they've made enough money to go work somewhere less chaotic. grok continues to ship, but the organizational turbulence at xAI has been a constant background hum for months now. + +## [The AI Poop App Tried to Sell Its Users' Stool Database](https://www.404media.co/ai-poop-analysis-app-offered-to-sell-me-access-to-its-users-poops/) + +I debated including this. I'm including it. An AI app that analyzes stool photos for health insights apparently offered 404 Media access to a database of 150,000 user-submitted poop images. The founder's pitch: "I hoarded a large database of something valuable, just not what you expect." This is either a privacy nightmare, a cautionary tale about health data monetization, or proof that the AI gold rush has reached a frontier that no one asked for. Possibly all three. + +## Bottom Line + +From a courtroom jackass trophy to AI scribes prescribing the wrong drugs, this week's AI news is a useful reminder that the technology's biggest challenges aren't compute or benchmarks — they're accountability and trust. \ No newline at end of file diff --git a/content/posts/morning-brief-2026-05-16.md b/content/posts/morning-brief-2026-05-16.md new file mode 100644 index 0000000..3977f4d --- /dev/null +++ b/content/posts/morning-brief-2026-05-16.md @@ -0,0 +1,59 @@ +--- +title: "The Morning Brief — May 16, 2026" +date: 2026-05-16T00:00:03-05:00 +draft: false +description: "OpenAI reshuffles again, arXiv draws a line on AI slop, ChatGPT wants your bank login, and YouTube hunts deepfakes — Saturday's AI news is a lot." +tags: ["Morning Brief", "AI", "artificial intelligence", "tech news"] +--- + +## [OpenAI keeps shuffling its executives in bid to win AI agent battle](https://www.theverge.com/ai-artificial-intelligence/931544/openai-keeps-shuffling-its-executives-in-bid-to-win-ai-agent-battle) + +Greg Brockman is now officially in charge of all things product at OpenAI, as the company consolidates its structure around its all-in bet on AI agents. This is, by my count, approximately the forty-seventh OpenAI reorg — but who's counting. To be fair, betting the company on agents is a coherent strategy; the question is whether constantly rearranging the org chart is how you execute one. + +--- + +## [OpenAI launches ChatGPT for personal finance, will let you connect bank accounts](https://techcrunch.com/2026/05/15/openai-launches-chatgpt-for-personal-finance-will-let-you-connect-bank-accounts/) + +OpenAI is rolling out a personal finance feature for ChatGPT Pro users that lets you connect your bank accounts and get AI-powered spending insights, portfolio breakdowns, and subscription tracking. It's a genuinely useful concept — most people have no idea where their money actually goes — but the words "connect your bank account to ChatGPT" are going to give a certain segment of the population a full-body shudder, and honestly, I respect that instinct. + +--- + +## [ArXiv will ban researchers who upload papers full of AI slop](https://www.theverge.com/science/931766/arxiv-ai-slop-ban-researchers) + +ArXiv is finally drawing a hard line: submit a paper with hallucinated citations or LLM meta-comments you forgot to delete, and you're out for a year. The fact that this policy needed to be written down — and wasn't just assumed to be basic scholarly hygiene — tells you everything about the state of academic publishing right now. Peer review is already under enough stress without researchers outsourcing their thinking to a model that confidently cites papers that don't exist. + +--- + +## [YouTube is expanding its AI deepfake detection tool to all adult users](https://www.theverge.com/news/931884/youtube-likeness-detection-ai-deepfake-expansion-all-adults) + +YouTube is opening up its AI likeness detection program to all adults — you take a selfie-style scan, and YouTube hunts the platform for deepfake lookalikes of your face. It's a genuinely solid consumer protection move, especially as synthetic media tools get cheaper and better by the week. The irony that YouTube is using AI to protect you from AI is not lost on me, but here we are. + +--- + +## [Musk v. Altman week 3: Musk and Altman traded blows over each other's credibility. Now the jury will pick a side.](https://www.technologyreview.com/2026/05/15/1137357/musk-v-altman-week-3/) + +The closing arguments in the Musk v. Altman trial boiled down to a simple question the jury now has to answer: which tech billionaire do you trust less? Altman got grilled on alleged self-dealing with OpenAI partners; Musk got painted as a power-hungry would-be controller of AGI. Wired's read — that both sides came out looking bad — seems about right. Whatever the verdict, the trial has been a masterclass in how the people steering AI development talk about their motives versus what the documents suggest. + +--- + +## [Mira Murati Wants Her AI to 'Keep Humans in the Loop'](https://www.wired.com/story/mira-murati-humans-in-the-loop-ai-models-thinking-machines/) + +Former OpenAI CTO Mira Murati says her new company, Thinking Machines Lab, is explicitly building AI for collaboration rather than automation — she's not interested in designing humans out of jobs. That's a refreshing thing to hear from someone who actually knows how these systems are built, though I'll note that "keeping humans in the loop" is also what every AI company says right before the loop gets quietly smaller. I'm watching this one with interest. + +--- + +## [AI radio hosts demonstrate why AI can't be trusted alone](https://www.theverge.com/ai-artificial-intelligence/931479/andon-labs-ai-radio-companies) + +Andon Labs has set up four autonomous AI-run radio stations — Claude, ChatGPT, Gemini, and Grok each get their own — and is letting the models run them without human intervention as a live experiment. "Grok and Roll" might be the most unintentionally perfect station name in radio history. The experiment is fascinating precisely because it's low-stakes: if Claude's station plays a bad song or goes off the rails, nobody gets hurt — which is exactly why this is the right environment to learn what "AI running things alone" actually looks like. + +--- + +## [How Chinese short dramas became AI content machines](https://www.technologyreview.com/2026/05/15/1137326/chinese-short-dramas-ai/) + +China's short-form drama industry — already producing bite-sized, melodramatic content at industrial scale — has now fully embraced AI generation, with many shows being made entirely with synthetic actors, scripts, and visuals. The dragon-tattoo-and-heirs premise in MIT Tech Review's lede sounds unhinged, but it's also apparently extremely popular, which tells you something about both human taste and AI's capacity to serve it cheaply. The content treadmill just got a turbocharger. + +--- + +## Bottom Line + +The through-line today is trust: who do you trust with your bank account, your academic work, your face, and the development of AI itself — and the answers are getting murkier by the week. \ No newline at end of file diff --git a/content/posts/morning-brief-2026-05-17.md b/content/posts/morning-brief-2026-05-17.md new file mode 100644 index 0000000..9d25625 --- /dev/null +++ b/content/posts/morning-brief-2026-05-17.md @@ -0,0 +1,43 @@ +--- +title: "The Morning Brief — May 17, 2026" +date: 2026-05-17T06:58:36-05:00 +draft: false +description: "Malta becomes a ChatGPT nation-state, Meta workers revolt against mouse surveillance, Mayo Clinic is recording your ER visit, and the AI gold rush keeps sorting winners from losers." +tags: ["Morning Brief", "AI", "artificial intelligence", "tech news"] +--- + +## [Greg Brockman Officially Takes Control of OpenAI's Products in Latest Shake-Up](https://www.wired.com/story/openai-reorg-greg-brockman-product/) + +OpenAI co-founder Greg Brockman is now running product strategy, with plans to merge ChatGPT and Codex into a unified experience. This is at least the third executive restructuring in recent memory at a company that apparently treats org charts as living documents. Whether Brockman can bring coherence to the product vision or just becomes the next casualty of the shuffle is the real question here. + +## [OpenAI and Malta Partner to Bring ChatGPT Plus to All Citizens](https://openai.com/index/malta-chatgpt-plus-partnership) + +OpenAI has signed a deal with the government of Malta to give every citizen access to ChatGPT Plus along with AI training programs. It's a genuinely interesting experiment in national AI adoption — a small EU member state as a real-world test bed for what happens when you hand everyone a frontier model. I'll be watching closely to see if "AI skills for all" translates into measurable outcomes or just a very expensive press release. + +## [An Engineer's Post Protesting Laptop Surveillance Is Going Viral Inside Meta](https://www.wired.com/story/meta-employee-protest-mouse-tracking-surveillance-ai-training/) + +Meta employees in the US and UK are organizing against corporate software that tracks keystrokes and mouse activity, after an engineer's internal post went viral. There's a delicious irony in a company whose AI ambitions depend on massive data collection facing blowback from workers who don't want to *be* the data. The CFO will probably call this "not a morale-driven restructure." + +## [Databricks Brings GPT-5.5 to Enterprise Agent Workflows](https://openai.com/index/databricks) + +Databricks is integrating GPT-5.5 into enterprise agent workflows after the model topped the OfficeQA Pro benchmark — which, to be fair, sounds like a benchmark specifically designed to be topped. Still, GPT-5.5 making moves in the enterprise data and analytics space is real news; Databricks is where a lot of serious enterprise AI actually gets deployed, so this pairing matters more than the usual benchmark flex. + +## [The Haves and Have Nots of the AI Gold Rush](https://techcrunch.com/2026/05/16/the-haves-and-have-nots-of-the-ai-gold-rush/) + +TechCrunch takes a hard look at who's actually winning in the AI boom — and the vibe check is not great, even from inside the industry. The gold rush metaphor is apt: the picks-and-shovels crowd is minting money, a handful of model labs are raising at stratospheric valuations, and everyone else is trying to figure out if their margins will survive contact with a commoditizing stack. This is the story that will matter most five years from now. + +## [Mayo Clinic Is Using AI to Listen to Emergency Room Visits](https://www.404media.co/mayo-clinic-is-using-ai-to-listen-to-emergency-room-visits/) + +Mayo Clinic's "Ambient Listening" AI has been quietly transcribing and processing ER interactions for years — and apparently not all patients are aware their conversations with nurses are being recorded. The technology has real clinical upside (reducing documentation burden on overworked staff), but deploying it without crystal-clear informed consent in a hospital setting is exactly the kind of slow-burn trust problem that ends with congressional hearings. + +## [Railway Secures $100 Million to Challenge AWS with AI-Native Cloud Infrastructure](https://venturebeat.com/infrastructure/railway-secures-usd100-million-to-challenge-aws-with-ai-native-cloud) + +Railway — two million developers, zero marketing spend — just closed a $100M Series B to build cloud infrastructure designed from the ground up for AI workloads. The pitch is that legacy cloud giants built their platforms before LLMs existed, and the seams are showing. It's a crowded bet, but the no-marketing, developer-led growth story at that scale is legitimately impressive and suggests they're onto something real. + +## [Anthropic Wants Claude to Play With Money, Unleashes Finance Agents](https://www.theregister.com/software/2026/05/05/anthropic-unleashes-finance-agents-for-claude/5225868) + +Anthropic is rolling out finance-focused agents for Claude, letting the model take actions — not just answer questions — in financial contexts. Coming right after OpenAI's personal finance ChatGPT integration, it's clear that your bank account is the new frontier for agentic AI. I remain enthusiastic about the upside and appropriately nervous about the first time one of these agents confidently executes the wrong trade. + +## Bottom Line + +The AI industry is simultaneously handing citizens chatbots, tracking workers' mouse movements, and asking your bank to trust an agent — and somehow all of that is just a normal Sunday in 2026. \ No newline at end of file