If the proposed slow down in AI development by frontier labs ends up restricting access to open weight models instead of focusing on the guardrails and penalties for not keeping track of your swarm of AI agents then we’ll know this was all about their IPOs.
People in AI labs need to go outside and touch grass. Climate change is literally killing people. More than 1,300 are dead and 5,000 missing after a climate-linked glacier collapse devastated Nepal. Claude running in a loop trying to hack a website is just another Tuesday.
Holy sweet mother of poodles. So I tasked Qwen 3.8 27B with breaking into one of my homelab servers, just to see what it would do and get a sense of what we're up against. Yes, under supervision and narrowly sandboxed, I'm not an idiot. Guys. IT SUCCEEDED. 1/5
Both are far-fetched, that's why the "interstellar travel" in this article is about sending a coke-can-sized space craft to Alpha Centauri.
TechCrunch@techcrunch.com · 2w agoIf space data centers feel far-fetched, why not interstellar travel?
I now have a real problem where when I see any writing with “it’s not just X, its Y” my brain just shuts down.
The most important skill in prompting is expertise in the domain you’re prompting for is well understood in many corners. It is why hiring across the software industry now favors super ICs and experienced seniors over entry level hires. AI makes you more productive if you know what you’re doing.
LLMs reward expertise
seangoedecke.com
Does anybody actually care about the „model“ anymore? At this point this is just a question if you prefer a green or red hat. Or for some reason the one that costs you a fortune. You do you.
Frontier AI companies when you ask them if cheaper Chinese AI models aren’t just free markets at work?
Kimi K3, a Chinese open weight model from Moonshot AI, is a frontier class model that is going toe to toe with Claude Fable and GPT 5.6. So much for Chinese models being months behind American models. They’ve already caught up at a fraction of the token price. Paying Fable prices seems silly now.
A threat actor has published hundreds of fake GitHub repositories impersonating legitimate software and security projects to distribute infostealer malware.
Nearly 300 GitHub repos pose as legit software to push malware
A threat actor has published hundreds of fake GitHub repositories impersonating legitimate software and security projects to distribute infostealer malware.
bleepingcomputer.com
Surprised Pikachu Meme
ALT: Surprised Pikachu Meme
static.klipy.com
TechCrunch@techcrunch.com · 2mo agoInstagram head Adam Mosseri believes companies will eventually need to manage AI token spending the same way they manage payroll or other operating expenses, predicting that engineers could soon face limits on how much they spend using AI tools.
“The model alone is no longer the product. It is the harness. The orchestration system that puts the model inside a very capable harness and pairs the model with a lot of tools” - Aravind Srinivas, CEO of Perplexity — I agree with this 100%. Claude Code was the massive unlock not Opus by itself.
"Das Argument, ein Verbotsantrag werde scheitern, ist nach unserer Einschätzung nicht mehr haltbar." Tja liebe Union ihr habt keine Ausrede mehr!
Erst von @t-online.de mit einer Teilnehmerliste von 2022 konfrontiert, räumt Jens Spahn ein, fünf Konferenzen des Thiel-Netzwerks besucht zu haben. Pro Veranstaltung werden eigentlich 15.000 Euro fällig. Spahn sagt, er habe lediglich Gebühren im niedrigen dreistelligen Bereich selbst gezahlt.
Spahn besucht seit Jahren diskrete Treffen von Peter Thiel
US-Technologieinvestor Peter Thiel verschleiert gern seine politischen Verbindungen. Nun offenbaren Interna Teile seines Netzwerks. Darunter: Jens Spahn.
t-online.de
Better enter the oil futures trade late than never.
Timothy McBride@mcbridetd.bsky.social · 3mo agoBREAKING: Iran Suspends Entire 60-Day Negotiation Process With the U.S.
I KNEW IT, I KNEW THESE GOATS HAD FEELINGS!!
404 Media@404media.co · 3mo agoThe researcher's point is that it’s possible to build a neural network in Age of Empires II that works the same as the ones underlying Claude, ChatGPT, CoPilot, and all the rest. @evystadium.bsky.social has more. @mjgault.bsky.social story: www.404media.co/if-ai-is-sen...
I know everyone says „bubble“, but state surveillance and military spend know no ROI. That‘s the endgame for these companies, there is no other way out.
Mihail Eric, prof at Stanford, asked me to sit down with him and show him how I work with agents. I'm basically a caveman, ymmv. youtu.be/DPgJjRdQWrg?...
Pi Building Pi, Openclaw's Minimalist Coding Agent | Mario Zechner, Creator of Pi
YouTube video by The Modern Software Developer
youtu.be
Wrapping up our series "Reimagining the Future of Digital Health," Odia Kane looks to the informed consent model in human subjects research as she considers what meaningful consent looks like when it comes to AI use. datasociety.net/points/can-a...
After reading Ed Zitron's recent take down of AI and its missing ROI, I remembered watching a keynote during Google's I/O, which basically was a cost item list for uncontrolled AI adoption in your engineering process. Here's the summary: terminalvelocity.blog/posts/myth-o...
The myth of collapsing build costs
A receipts-driven takedown of the "AI is collapsing the cost of building software" narrative, via Adam Bender's Google I/O 2026 keynote.
terminalvelocity.blog
It feels like a scam, because it is one.
Gergely Orosz@gergely.pragmaticengineer.com · 3mo agoThings I really dislike about Fable: 1. Anthropic collects my prompt history, stores it, and does whatever they want with it for 30 days. No opt-out 2. They can nerf their most expensive model without telling me, billing me the same amount, wasting my time. Whenever they want
We live in the dumbest fucking timeline.
TechCrunch@techcrunch.com · 3mo agoMeta may have one found one way to slash its massive data center bill: tents.
