Search: “agents”

4⃣4⃣4⃣4⃣ Introducing DeepSeek V4 Pro, a long-context model with hybrid attention, three reasoning modes, and SOTA coding performance. AI natives can now use DeepSeek V4 Pro on Together AI and benefit

Cloud-hosted models will become more niche/seldom-used as much cheaper local models fulfill that which 90% of consumers need from AI. Lots of models to buy/use, but very few hardware options upon whic

Congrats to @deepseek_ai team! Doing the numbers I would estimate: Pro < 14m for the final training run Flash < 4m Ratio of active params x total training tokens vs v3 Total compute costs (data prep,

For the last 72 hours since ml-intern launched we have had over 500+ autonomous AI research projects running on the Space at all times. Some insane ones I saw: 1. A new AI paradigm from scratch — tryi

Introducing DeepSeek V4 Pro, a long-context model with hybrid attention, three reasoning modes, and SOTA coding performance. AI natives can now use DeepSeek V4 Pro on Together AI and benefit from reli

Just merged a built-in skill for Google's DESIGN.md A skill that lets Hermes author, lint, diff, and export DESIGN.md files, giving it fluency in Google's new open-source visual-identity format the mo

Models are changing really fast, often in ways that will break your product and technical assumptions. There's no way around this – the only way forward is to relentlessly iterate to make sure you're

CursorBench is a benchmarking tool or evaluation framework developed by Cursor, likely designed to measure and compare the performance of AI code completion or development features. The post provides

New episode of The Information Bottleneck is out, this time with @liuzhuang1234 (Princeton). We talked about ConvNeXt and whether architecture still matters; dataset bias and what 'good data' actually

This has been the most anticipated event in OSS and it doesn't disappoint. Congratulations to @deepseek_ai team. DSV4 Pro is now available on @togethercompute and we'll be adding a lot of capacity beh

At #RenderCon2026, @EMostaque makes the case during How to Build the Holodeck that stories still matter, IP still matters, but the barriers to building entirely new experiences are dropping faster tha

GPT-5.5 in Codex is a delight to work with: - Super sharp with responses - It understands intent better than any model - Great 'personality' - Gets lots of stuff done without pausing unnecessarily It

I've been previewing this in Codex for a few weeks - it's very good! Had some great results from it having it run security reviews against code written using other models Introducing GPT-5.5 A new cla

looks like new Pareto frontiers across everything: - Context: 400K context in Codex and a 1M in API - API Pricing: 5/m input and 30/m output tokens. - Codex improved its own inference speed 20% lol -

🚨 OpenAI just launched GPT-5.5. The OpenAI team was nice enough to give me early access over the last several weeks, and I just want to flag: there is a certain class of models (one that we’re hitting

Things have been degrading super fast in Claude Code. I still use Claude Code, but my default is now Codex. I still prefer Opus models for coding, and so I will try again with the fixes. I appreciate

Boris Cherny announced that usage limits are being reset for subscribers in response to user feedback and patience. The post suggests this is a policy adjustment made by a service or platform to addre

This is a really big deal - it's easy to run into nasty bills with Cloud Run if your site attracts aggressive scrapers, spending caps make it a whole lot safer to run small projects on Anouncing Spend

I built a web analytics system to track traffic and events across all my @Replit projects in one place. Drop-in @posthog stack: → Full analytics dashboard → Reverse proxy (beat ad blockers) → AI skill

I find that open weights models over-perform on benchmarks compared to actual real-world usage, and Kimi feels like no exception. For example, a small amount of use will show that Kimi is not as good

K2.6 + hermes = 4 hr session setting up qwen 3.6 training regime on dgx spark with the current autnomous session currently lasting 70+ min without any prompting. Kimi with hermes is next level. @NousR

This post likely discusses how the next generation of team communication and collaboration platforms will diverge from Slack's current design paradigm and feature set, possibly exploring new UI/UX app

Cursor allows users to customize the CLI status bar through the /statusline command to display personalized session and runtime information. This feature enables developers to configure which signals

Earlier this year Yann LeCun left Meta because Mark Zuckerberg wouldn't bet the company on JEPA. Last week his group dropped the first JEPA that actually trains end-to-end from raw pixels. 15 million

Excited to go public with Logos, a first principles system designed for intuition & innovation We are starting by sharing results in physics before moving to other domains Today’s is a lot of fun, one

Free ~20-50% tok/s on a local llama.cpp setup if you already have a draft model sharing vocabulary with your main one. Local stack quietly got faster this weekend https://x.com/TechIno219886/status/20

.@gradypb opening @sequoia AI Ascent event. This is AGI! advice for founders competing w/ the labs for the $10T of market opportunity: 1. get MAD (Moats, Affordances, Distribution) 2. approach every p

Simon Willison shared an experience attempting to use X's Community Notes feature to add clarification to his own tweet, experimenting with whether the self-noting functionality would work as intended

Kimi K2.6 just dropped. And it crushed Claude Opus 4.6 on SWE-Bench Pro. Kimi K2.6: 58.6 GPT-5.4 xhigh: 57.7 Gemini 3.1 Pro: 54.2 Claude Opus 4.6: 53.4 An open source Chinese model is now #1 on agenti

Time to really give Kimi K2.6 a go. Thank you @ollama! Love the ollama cloud setup! Kimi K2.6 raises the bar for open-source models. 🦙 available on Ollama's cloud! Try it with OpenClaw: ollama launch

🔴 BREAKING Speculative decoding just got real for local LLMs. llama.cpp merged speculative checkpointing (PR #19493). Expect 0-50% speedups on coding tasks with `--spec-type ngram-mod`. This means: •

I am not convinced that we should be comfortable calling 'problem solving' or 'judgement' or whatever as skills that are impossible for AI to do well. Like any other skill, there are humans who are re

This post is a thank you acknowledgment from the Ollama project to a contributor (@scaryrawr) for their assistance with pull request #15583 on the Ollama GitHub repository. The tweet references a spec

That didn't take long for a total U turn. First the Nvidia CEO boasted AI would take a vast number of jobs. But now that there's a huge backlash against AI, he cobbles together this cope - to try to p

Anthropic’s CEO keeps talking about AI wiping out jobs because he’s trying to IPO this year. If he positions Claude as armageddon for jobs, his TAM becomes “all white-collar human labor,” not just AI

I sat down with Emad Mostaque (@EMostaque)—and he told me something I can’t shake: the biggest AI labs already have models they’ll never release. We dug into why, how reinforcement learning may be kil

I think Artificial Analysis does a good job overall and provides transparency in benchmarking, but GDPval-AA is not a good benchmark and needs to stop being reported. It is Gemini 3.1 judging other mo

Okay this one is insane. A new 18B frankenstein model was just released on @huggingface — Beats the new Qwen3.6-35B-A3B on a 44-test suite despite requiring 12GB VRAM instead of 24GB 🤯 Runs on a SINGL

in retrospect putting the slop cannons (@_lopopolo) on @aiDotEngineer talks day 1 and putting the grown ups (@badlogicgames) on talks day 2 is working out pretty well for faithfully representing the m

In @steipete's latest State of the Claw, he gives an update on 5 months of @OpenClaw and some behind the scenes on what it's like maintaining the fastest growing open source of all time: https://www.y

Oh look! Anthropic's entire 'we are delaying Mythos' narrative was marketing hogwash. Kudos to FT for confirming what was obvious. Anthropic simply doesn't have the compute. FT: 'Multiple people with

Claude Code's session management capabilities and 1 million token context window are highlighted in this post, which references an official Anthropic blog entry. The feature allows developers to maint

I edited the intro because I realized I buried the lede originally- The 1M context window is a double-edged sword. It allows Claude to do more complex tasks but it can also leads to more context pollu

I love to explore solutions with it so I often ask it to brainstorm with me, then choose an option and rewind to implement it. When I'm ready, I ask it to interview me to figure out what’s in my head

Replit shared a YouTube video detailing their company's journey and history of development. The video likely covers key milestones, founding story, product evolution, and the team's vision for their c

🎬 Ollama Gemma Day Recap: SGLang at the Ollama Gemma 4 Party in Palo Alto 🍾 Last night, @ollama hosted a packed Gemma Day at the Palo Alto office alongside the @GoogleDeepMind Gemma team. SGLang was i