Freeing the Law with LOCUS A Local Ordinance Corpus for the United States https://t.co/SmLiqU9Ziy

paper: https://t.co/XTeC0IXcyI

@LLMathematician I used the ZAI api directly https://t.co/WlVTEHU2Jv

1% of people account for 63% of all violent crimes. 0.2% of people ever commit murder, and **67% of all murders**are committed by people with prior arrests You can literally just fix crime by not tolerating people who show a history of being destructive to society. https://t.co/AD0675q5NV

// Evolving Meta-Skill for Multi-Agent Systems // Can a multi-agent system get better at orchestration without touching a single weight? Automatic MAS generation has been stuck between two bad options. Inference-time methods use frozen frontier models but never learn from past runs. Training-time methods learn but are capped by small-model capability. Skill-MAS takes a third path. It treats the orchestration capability as an evolvable Meta-Skill, refined through a closed loop of multi-trajectory rollout and selective reflection that distills experience into strategy-level principles rather than memorized traces. Across four benchmarks and four different LLMs, the evolved Meta-Skills transfer to unseen tasks and other models, because the know-how lives in text at the strategy level, not in any one model's weights. You keep the frontier model and still accumulate experience. Paper: https://t.co/fn4J2Gz33M Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

Great paper, made it available here: https://t.co/pwvKtvpzLq Check how it compares to other text-to-image models at the bottom https://t.co/ODCM9RUMYn

Great paper, made it available here: https://t.co/pwvKtvpzLq Check how it compares to other text-to-image models at the bottom https://t.co/ODCM9RUMYn

Here's how to use Claude Code with GLM-5.2 via @huggingface Inference Providers: 1. Create a token at https://t.co/K7LA8W4aWh (fine-grained, enable "Make calls to Inference Providers") 2. Set the environment variables below 3. ??? 4. Profit https://t.co/mnZmXWWsRP

New paper: every law in America is technically public. But not really, until now! With @DenisPeskoff at UC Berkeley, we built a corpus of ~every publicly accessibly city and county law, and released a huge chunk of it! 2.2 million laws, you're (probably) covered in it! 🧵

You can do this with a Steam Controller. https://t.co/7bGc3NWiVz

@yunta_tsai 💯 https://t.co/rIoRkl86Jh

@yunta_tsai 💯 https://t.co/rIoRkl86Jh

THE AGENT THAT GROWS WITH YOU 💫 @Teknium @NousResearch https://t.co/HfcnvykjRa

THE AGENT THAT GROWS WITH YOU 💫 @Teknium @NousResearch https://t.co/HfcnvykjRa

Life lately I’m headed back to San Francisco with. New lease on on life (my nose works now) and a new friend in New York w/ @karmicdebt Landing in 5. Pool party in Tiburon. Exa series C party 12 day work week. stoked to come on some pods this summer. Talk about everything I’ve learned. Stoked to see everyone at @swyx ‘s conference. Excited to see @thsottiaux flourish in as fearless leader of the singularity Excited to see so many new friends on @romainhuet team

+2 more

GLM-5.2 is great at design (Opus level IMO). I am also starting to see great results with long-running tasks, too. How is this possible? I think there are a few clever hacks. But I just came across this from the official blog, and they actually trained this model with an anti-hacking module. RL, as many know, comes with this issue of reward hacking that often enables the model to take weird and suboptimal shortcuts. Not only that, but it makes the models sometimes feel like it's sometimes "lazy" or just plain "dumb" at times, including other issues like intent misalignment, verbosity, sycophancy, deception, etc. And you really don't want that for long-running tasks operated by coding agents. This is a great insight. If you use the standard /goal (in 5.5 or 4.8), you notice the models often take shortcuts that lead to long-running tasks (wasting tokens along the way) but with poor results. This is why I advocate for a focus on better verifiers. So this anti-hacking idea is a model capability that should, in theory, lead to better results on long-horizon tasks. I've seen efforts here and there in a few research papers, but haven't seen it translated to much, much less in a frontier, open-weight model. This might be contributing to some of the great results we are seeing with GLM-5.2, but I suspect there is more, of course, like better verification capabilities. It's not clear how all of these training signals lead to downstream capabilities, but this is something to look at closely with newer models.

GLM-5.2 has been "stuck" at No.2 on Hugging Face Trending for three days, but I'm thrilled to have connected with the creator behind the No.1 project this afternoon. It's been amazing to see open-source work resonating with so many people. https://t.co/fRb4UXvtvV

New research: Is AI making employers view labor as more of a commodity? In a large online labor market, we find that post-ChatGPT, clients place less weight on signals of human capital and more on price when hiring. https://t.co/5U0ObQsI6Z

I bet you couldn’t post your chats. https://t.co/NUdBqNtQZC

@archived_videos https://t.co/oP0qShZqIy

*Another* apparently AI-generated story wins a literary prize, this time judged by a panel including the novelist Ruth Ozeki. Literary prizes need to start including Pangram checks in their process, or else change the rules to make AI writing ok. It’s very simple! https://t.co/Aw3MYR9xMZ

+1 more

Good lord this edit. Insane. https://t.co/gLlLFszugs

S-Agent Spatial Tool-Use Elicits Reasoning for Spatial Intelligence https://t.co/sbFDXYe1ZX

paper: https://t.co/ljt60z8zjn

American media would never lead with my realism about AI. Fun to see other parts of the world less cowed by the tech industry. https://t.co/i02kyiO7x3

"If Bores wins the primary, it has the potential to clarify and perhaps even transform the federal debate about regulating artificial intelligence. On the other hand, if he loses, Bores and others fear that cautious Democrats—in New York and nationally—could be even less likely to stand up to the AI tech billionaires." We can't let the oligarchs win. https://t.co/sCpaJ9FsUo

Doctors 25% more likely to miss a diagnosis on a patient within 3 months after Ai is implemented. The results indicate Ai has an immediate negative effect on professionals skillsets vs prolonged reliance. https://t.co/ciPB3DdR2J

Goldman Sachs is now saying the AI race has become a $5.3T capital-spending cycle. with that figure covering expected hyperscaler spending on AI and data centers from 2025 through 2030. AI infrastructure is starting to strain normal financing channels, because the same few hyperscalers cannot endlessly push debt into public bond markets without investors worrying about issuer concentration. A data center is not one asset, because it combines land, power access, network links, buildings, cooling, and AI servers, so the financing naturally spills across infrastructure funds, real estate funds, private credit, and corporate bonds. Goldman signals that AI capex estimates are rising faster than actual data center construction, which means the bottleneck may shift from model demand to financing capacity, power availability, and project execution.

@matvelloso What AI bubble? This is fine. Just spawn +100 agents to verify the security leaks on the previous ones. https://t.co/pAxIe7Y61J

@TheStalwart @tracyalloway @jackclarkSF @PeterMcCrory https://t.co/2bTzmFEjPw

@leilavclark https://t.co/2bTzmFEjPw