we need to talk about how GPT-5.6 Luna beats Opus 4.8 at agentic coding, with a fraction of the cost. and it's not even far off from Fable 5. now this is on benchmarks, so take that with a pinch of salt. but the majority of coding tasks aren't much harder than bench-level.

GPT 5.6 beats Fable 5 by a margin on a research-level physics problem. even older GPT models perform better on this unpublished benchmark. Grok 4.5 however, performs poorly with a score of 15.4%.

AI 2040 was released today by the team that wrote the AI 2027 report. the interesting part: they recommend AGAINST releasing frontier open models publicly. "terrorists and other actors might be able to use highly capable open weight models to design world-ending bioweapons" regardless of their take, I think it's inevitable for open source AI and especially Hugging Face to go through some deep troubles in the next few years, as open models are not insignificant to be ignored by policymakers anymore.

so you're saying just by getting an X Premium+ sub, you get: > Grok Imagine for images/videos > SuperGrok for general use > Grok Build + Grok 4.5, which is Opus level and cheaper in itself > AND monetize X to earn back the money is that right @elonmusk? insane ROI

these guys are making Claude, Hermes, and DeepSeek argue with each other instead of just agreeing with you. this is the practical future of agent teams. with @Bloome_im you just pick a task template, and it spins up a full agent team in a group chat to collaborate and get the job done. research, review, writing, market analysis, or anything you want. they brainstorm, review, and plan together. meanwhile, you or other human teammates can steer them or jump in, and stay in the same conversation so nothing gets lost across tools. since this is actual work, Bloome works across web, mobile, and desktop. try the desktop app first if you're using it seriously. same flexibility on the agent side too, you can connect any agents you want: ChatGPT, OpenClaw, Gemini… check them out →

Anthropic asked big pharma for their data, everyone said NO if you’re wondering why AI won over software but hasn’t made major discoveries yet, this is one main reason. it’s not just that AI isn’t smart enough. but big pharma realizes their data is their moat, and aren’t willing to hand it over to anyone. and probably as we go along, more companies will realize this and start OWNING THEIR WEIGHTS by going for open source options.

Grok 4.5 is cheap af. it's Opus-level frontier intelligence at ~25% of its price, just a little above the price range of Chinese open-source models. I've never been as impressed by Grok. let's hope it holds up in practical use.

Elon wasn’t messing around. Grok 4.5 is indeed Opus level, faster, and cheaper. Never thought I’d live to see this day but OpenAI and Anthropic have some serious competition now.

OpenAI’s new GPT-live fixes some of the most annoying problems with their previous Voice Mode: > it can listen and speak at the same time > better turn taking, knowing when to speak > delegates hard work (reasoning, search, plan) in the background > keeps talking while reasoning > better listening with background noise > stronger on benchmarks OpenAI already had a big lead in this but with the new generation the gap is gonna be huge.