But... what about our jobs?
Last week's main discussion was about jobs, which unsurprisingly lead to this week being about... jobs (and careers, and corporate adoption.)
@stack_harder asked: Is college still worth it in the era of AI? and stackers in majority countered that higher education has been a mirage for a longer time and that this isn't an AI phenomenon, yet @Scoresby shared that Generative AI is reducing the number of junior people hired. Is there a chance stackers are just wrong? @kepford shared The Blue-Collar Boom: AI's Impact on the Workforce. But this in turn was weakened again by @k00b's share: US manufacturing has lost 42,000 jobs since April. Can the world make up its mind please? Uncertainty all over the place!
@hn reposted You don't want to hire "the best engineers", which makes sense, because Zuck bought them all a couple of weeks ago with up to 100M shittokens signing bonuses. Should AI Get Legal Rights? was shared by @0xbitcoiner, which would be awful: that would allow the #@%! bot that took your job to unionize. You'll never have a job again!
Doomers may rule the narrative outside SN, but there was good news on the topic too: OpenAI is building an AI jobs platform that could challenge Microsoft’s LinkedIn, @Coinsreporter shared. This is great because now we can segregate job offers: AI's go to OpenAI's job platform, and human career chasers just stay on LinkedIn. Also, AI adoption actually falling now for many firms was shared by @zuspotirko.
Good news too for all the former thought police that were fired at Twitter and Meta: OpenAI needs you! OpenAI employees review your chats and may refer to law enforcement, @Scoresby shared. Isn't that the perfect job? You can now get paid to do deep privacy invasion and be a snitch at the same time! I for one will make sure to make build some completely inappropriate language in my code bots.
More questions:
- Would you trust an AI doctor to diagnose you? by @Kayzone
- What’s one big weakness of AI that doesn’t get talked about enough? by @Kayzone
- What’s the scariest thing you think AI could do in the next 5 years? by @Kayzone
- How do you spot the difference between AI-generated content and human work? by @Kayzone
More Friction:
- Why do We Still Play Chess? by @TotallyHumanWriter
- Tesla’s new ‘Master Plan’ sounds like AI slop by @0xbitcoiner
- From the distributed dream to the digital feedlot by @roistrdn
- Why can’t chatgpt look up current BTC-USD by @satgoob
- ART AND ARTIFICIAL INTELLIGENCE: THE END OF HUMAN CREATIVITY? by @Msd0457890
Sources?
There was a side-discussion about what sources are used by AI: 10.4% of AI Overview Citations are AI-Generated was shared by @0xbitcoiner and @guerratotal shared a nice graph depicting Where AI gets its facts says it all. @realBitcoinDog taught ChatGPT all of ~AI only to catch it red handed in anecdote theft in ChatGPT - “If there’s an SOP you’re SOL.”, except it turned out that the bot stole a stolen anecdote.
@Scoresby shared the news that AI startup Anthropic agrees to pay $1.5bn to settle book piracy lawsuit, which is fine, because there is now a benchmark for Challenging LLMs with Interpreting Nonsense with Depth, shared by @optimism, so we can just feed LLMs slop from now on and then they'll eventually get to profound insight, just like a sarcastic stacker would.
Explainers and guides
- ‘World Models’ an Old Idea in AI Mount a Comeback by @0xbitcoiner
- These psychological tricks can get LLMs to respond to “forbidden” prompts by @RideandSmile
- New AI Paradigm?! Energy-Based Transformers Explained by @carter
- How big are our embeddings now and why? by @carter
- How to spot fake AI by @redread
- Understanding Transformers Using A Minimal Example by @carter
- Don’t Build Multi-Agents by @carter
- The maths you need to start understanding LLMs by @hn
Models
- VibeVoice: A Frontier Open-Source Text-to-Speech Model by @carter
- The flagship model Qwen3-Max-Preview has been released by @lunin
- Hermes 4 outperforms OpenAI models with minimal content restrictions by @lunin
- Tencent Open Sourced a 3D World Model by @hn
Research
- Boffins craft certified way for AI to unlearn private data by @0xbitcoiner
- AI just Broke Trackmania's most Legendary Record by @jakoyoh629
- CauseNet: Towards a Causality Graph Extracted from the Web by @carter
- Inverse IFEval: Unlearn Training Conventions to Follow Real Instructions? by @optimism
- Communication Efficient LLM Pre-training with SparseLoCo by @carter
- A.S.E: A Repository-Level Benchmark for Evaluating Security in AI-Generated Code by @optimism
- 93% of GPT-4 performance at 1/4 cost: LLM routing with weak bandit feedback by @hn
Tools
- Maple Proxy – the Maple AI API that brings encrypted LLMs to your OpenAI apps by @0xbitcoiner
- Genspark has released an AI editor for video editing by @lunin
- Prototyper - Create apps in minutes by @deSign_r
- Le Chat. Custom MCP Connectors. Memories by @hn
- Claude Code: Now in Beta in Zed by @hn
Application and adoption
- Brian Armstrong says about 40% of Coinbase’s daily code is AI-generated by @0xbitcoiner
- As Xi and Putin chase immortality ponder digital dictators by @0xbitcoiner
- Simon Willison has vibe-coded 124 useful tools by @byzantine
- AI is going great for the blind. by @carter
- Qwen3 30B A3B Hits 13 token/s on 4xRaspberry Pi 5 by @hn
- Using Claude Code SDK to reduce E2E test time by @hn
Reports & Opinion
- Where's the Shovelware? Why AI Coding Claims Don't Add Up - Mike Judge by @Scoresby
- AI Agents Are Predicted to Increase the Software Market and Outstrip SaaS by @k00b
- The Alignment Battle We Already Lost by @antic
- ChatGPT’s new branching feature is a good reminder that chatbots aren’t people by @jakoyoh629
- Protecting the "human heartbeat" in the age of AI-led VFX by @redread
- How AI Actually Revolutionized the World by @anon
- Parallel AI Agents Are a Game Changer by @carter
- Why AI isn’t replacing engineers—it’s making them more creative by @NovaRift
- Almost anything you give sustained attention to will begin to loop on itself by @hn
- A staff engineer's 6-week journey with Claude Code by @carter
Safety & Security
- Detecting and countering misuse of AI by @carter
- OpenAI Announces ChatGPT Parental Controls by @Coinsreporter
- Detecting Exposed LLM Servers: A Shodan Case Study on Ollama by @RideandSmile
Government & Degens
- I’m a gamer. The NO FAKES Act could get me in trouble (reason) by @optimism
- China’s social media sites rush to abide by AI-generated content labelling law by @lunin
- How to Achieve Independence Privacy and Trust in the Adoption of AI by @0xbitcoiner
- How to Use Pump Parade’s AI Crypto Price Prediction Tool in 4 Simple Steps by @Kayzone
- Meet the Guys Betting Big on AI Gambling Agents by @0xbitcoiner
Big Tech
- 33 AI USA startups have raised $100M or more in 2025 - Tech Crunch post by @Car
- Why language models hallucinate - OpenAI by @Scoresby
- The SaaS in AI’s step by @BlokchainB
- OpenAI Boosts Size of Secondary Share Sale to $10.3 billion by @Cje95
- OpenAI Plans India Data Center in Major Stargate Expansion by @Coinsreporter
- OpenAI links up with Broadcom to produce its own AI chips by @Coinsreporter (#1209368 by @redread)
- Artificial Intelligencer: The private IPOs that fuel the AI boom by @NovaRift
- AI Week 2025: Recap by @0xbitcoiner
- Anthropic raises $13B Series F at $183B post-money valuation by @carter
- at $183B valuation, Anthropic AI has ~10% of the marketcap of bitcoin by @standardcrypto
- CoreWeave Raises 35,946 BTC For Pennsylvania Data Center by @BlokchainB