All entries

The fall of Rome was less a dramatic collapse and more the slow and eventually total decline of state services. One day, your postman becomes a little less reliable. He arrives less often. Then later

The hard part of reinforcement learning on a frontier model is the infrastructure that keeps training and inference numerically identical: zero KLD, end to end. We've solved this challenge, and are no

The Kimi API is now live on AWS Marketplace. 🚀 If your team is already running on AWS, you can now access Kimi with consolidated billing. Plus, eligible customers can apply Kimi API usage directly tow

Elon Musk posted about improvements to the X Money financial service experience, highlighting enhancements to card activation processes and expressing optimism about expanding the user onboarding. The

The most complex phenomena arise from scalable recombination of very simple rules. Whether it's galaxies, chips, or neural networks, if you find the right primitive building blocks, the complexity tak

The next scaling law is multi-agent swarms. Mixture of models. Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API. Our ‘Fugu Ultra’ model matches the pe

The problem we face with AI today is not a technical problem, it is a political problem. Of who gets to decide. What level of risk the public is exposed to, what future we build towards or avert. It's

The strange thing about progress is that it only counts if you ignore where we started. Every civilisation practised slavery. The civilisation that ended it is now told it should feel guilty forever f

The users who complain about the flaws in your product may seem annoying, but they are on the whole probably your most valuable users. They complain because they care, and I doubt a startup could ever

Clem Delangue discusses a resurgence of interest in decentralized intelligence systems and approaches, suggesting a shift away from centralized AI models toward more distributed architectures. The pos

Gary Marcus argues that generative AI, while significant, represents only a narrow subset of the broader computing field and should not overshadow the diverse and impactful work occurring across other

Thinking about AI Engineer World's Fair? Apply to be an Associate! It's a great way to connect with fascinating people across the AI ecosystem, build new relationships, and be part of one of the most

This is one of those moments where I feel most disconnected from my normie friends I'm like 'do you realize what this means?' And they say 'no' and I realize I will need to explain decades of nerd lor

This is the strongest ARC-AGI-2 performance to date by an open-source model. GLM-5.2 from @Zai_org on ARC-AGI (Verified) - ARC-AGI-2: 22.8%, 0.25 - ARC-AGI-1: 77.0%, 0.19 Performance is comparable wit

Yohei Nakajima discussed Activegraph at an AI House event, likely covering applications or insights related to graph-based AI systems or knowledge representation. This appears to be a social media pos

Nous Research recommends using the 'hermes skills config' command to enable creative-ideation and other optional skills in their system. This command appears to be a configuration tool for activating

Today I present you a curated collection of @tszzl saying that Fukuyama is always right We Need an International Treaty to Ban Superintelligence https://open.substack.com/pub/persuasion1/p/time-for-a-

Tracing functionality can be utilized to monitor and analyze memory usage patterns in applications, helping developers identify memory leaks, optimize allocation, and understand runtime behavior. This

Traitor-maxxing August, 2021. Mayorkas goes to the Southern border and announces a bunch of USAID funded programs to improve the lives of the people in Guatemala, Honduras, and El Salvador. We later f

Try Grok I switched the Grok as my search engine a few months ago. I only use Google now to correct typos / check spelling, as it's slower to use Grok for those tasks. Grok is superior to Google in ev

Try Kimi K2.7 and GLM 5.2 for free in Devin Desktop and CLI Kimi K2.7 Code and GLM 5.2 are available in Devin Desktop and CLI Both perform strongly on FrontierCode Extended, our benchmark for real-wor

Unlimited OCR is a great model on table parsing and understanding proper reading order. However it does struggle a little on semantic formatting, charts (it does decent at bounding boxes). Attaching t

USAID funded the Wuhan lab Wuhan lab created SARS-CoV-2 Wuhan lab was making more viruses DOGE defunded USAID DOGE may have been the most important pandemic prevention agency of the past century This

The MongoDB plugin in Grok Build enables users to query data, optimize database indexes, and perform database management tasks directly within the Grok interface. This integration allows developers to

very proud to host this session opening night! if you work in tech but have never heard of Touchy Feely, you are exactly the kind of person that needs to see this. You have no idea the shocking number

Watching Kyra and her family fight brain cancer — glioblastoma — over the last year has been inspiring, and heartrending, beyond belief. Now her endlessly proud father, @DavidPSamson, will continue th

We benchmarked Mistral OCR against other frontier and open-weight models on ParseBench 📊 For a model at its price point, it is quite competitive! - It wins on semantic formatting - understanding strik

We built Claude for outbound sellers. AEs & SDRs can harness GTM engineering through chat across 40+ data sources, no technical skills required. We’ve had 57,548 queries in our first few weeks of beta

We have a new version of GPT-5.5 Instant for you, and it's much more fun to talk to. Our most-used model is now better at understanding the intent behind a question and adapting its response according

We open-source Qwen-AgentWorld-35B-A3B (MoE, 35B/3B active, 256K context) and AgentWorldBench. Two routes, one roadmap: 🔬 Build the simulator — scalable, controllable, surpassing real environments 🧠 I

This post likely references Gary Marcus's critical commentary on limitations or failures in generative AI systems, using 'Fizzle' as a metaphorical or literal descriptor for underperformance or unmet

We’ve designed and built our first AI chip: Jalapeño. Designed from the ground up by OpenAI and brought to production with @Broadcom, Jalapeño is purpose-built for the LLM workloads powering ChatGPT,

We've provided some updated results on Mistral OCR that make use of the annotation feature for charts. The overall score is ahead of GPT-5.5 and just behind Gemini 3.1 Pro, which is quite impressive f

What is actually limiting model routers for coding tasks? Most routers treat picking a model as a static, one-off classification. This paper identifies the real bottleneck as information deficit. Simp

This post documents a Wordle game result where the player solved puzzle #1,830 in 3 attempts, with the final answer being a five-letter word where the 4th and 5th letters were guessed correctly in ear

This workflow template demonstrates how to use the Seedance 2.0 R2V (Rank-to-Video) model within ComfyUI, likely enabling users to generate or process video content based on seed values and ranking pa

Wow even @MicahLasher thinks the OpenAI-Andreessen-Palantir money spent to oppose @AlexBores backfired -- NY-12 retrospective in the NYT from Teddy Schleifer. Hard to imagine this was the headline tha

You can now delegate tasks to Cursor directly from Notion. It's built on the Cursor SDK, so every cloud agent runs on the same models, harness, and runtime that power Cursor. @​Cursor on any spec or a

Your Hermes Agent can now adopt an animated pet: a small sprite that reacts to what the agent is doing (idle, running a tool, thinking, waiting, finishing, failing) in the GUI or TUI. You have nearly

460 criminal charges 6.5B intended loss, 1.4B actual loss $127M seized assets Great work by @SecKennedy, @DAGToddBlanche, @FBIDirectorKash, & @DrOzCMS on today’s fraud takedown. If you’re committing f

A Biden-appointed activist judge, Sparkle Sooknanan, just blocked a commonsense tool to keep non-citizens from voting. The SAVE system simply encourages states to work with DHS to check their voter ro

A case study in why organizations should both incentivized their employees to explore AI uses that help them & have a Lab of dedicated AI builders Here, Cornell's finance & AI teams created a /treasur

Nous Research announced that their Hermes Agent has achieved a trillion tokens processed daily and reached 200,000 GitHub stars, reflecting significant adoption and performance milestones for their AI

ActiveGraph: 1 month in: 📄Paper #1: The Log is the Agent 🧠3 LongMemEval Experiments 🔄 Paper #2: Regimes, self-improvement loop 🎓 http://learn.activegraph.ai 🗂️ 2 reference agents (code, research) 💾 co

Gary Marcus argues that data security is fundamentally fragile and that mass surveillance and data harvesting practices create persistent security vulnerabilities that cannot be eliminated. The post e

AI Capex pushback from GS: 'If frontier intelligence can increasingly be developed in the East at a fraction of the cost incurred in the West … then the largest capital allocators are also the ones mo