samir

@icanvardar remember we're talking about a company with 400b revenue and 160b profit

@icanvardar they just have different priorities if they wanted do they could release the best coding model by end of the year

google should just give up on ai at this point

google finally started working on improving coding and agentic performance of their models theinformation.com/articles/googl…

"don't retweet this, don't retweet this, don't retweet this..." ah fuck it, life imitates art.

deepseek v4 is crazy cheap

@XFreeze bruh

I tested the same prompt on both Grok 4.3 and GPT 5.5: “Count to 10 starting from 11” ChatGPT 5.5 gave the obvious 11–20 Grok 4.3 gave 11, 10 and explained why going backwards was the only logical move Grok’s logical reasoning is at a level most models still can’t even touch

when banks collapse because of their own greed and bad bets, governments bail them out by printing money and dumping the cost on everyone when defi protocols suffer the ripple effects of a hack, they reach into their own pockets to protect their users this is the world i want defi united

Aave service providers have published an Aave DAO governance proposal to contribute 25,000 ETH to the ongoing DeFi United effort. Aave DAO's ETH contribution would go towards the plan to restore rsETH's backing to try and normalize market conditions as quickly as possible. DeFi United. governance.aave.com/t/arfc-rseth-i…

JUST IN: Google to invest up to $40 billion in Anthropic as part of a deal that includes 5+ gigawatts of computing capacity over five years.

google is still tying their shoes in the ai race

GPT 5.5 just refactored my entire codebase in one call🤯 ◈ 67 tool invocations. ◈ 4,200+ new lines. ◈ 96 brand new files. ◈ It modularized everything. ◈ Broke up monoliths. ◈ Cleaned up spaghetti. None of it worked. But boy was it beautiful.

@0xSero he's right tho

The models generally pleasant to talk to compared to other GPT models, kinda funny. Spuddy you might say (Image from Kitze)

Over the past month, some of you reported Claude Code's quality had slipped. We investigated, and published a post-mortem on the three issues we found. All are fixed in v2.1.116+ and we’ve reset usage limits for all subscribers.

@ClaudeDevs i just asked claude code to tell me a joke and i spent 2% of usage, you're lying

@OpenAI "our most powerful model yet"

GPT-5.5 delivers this step up in intelligence without compromising on speed. GPT-5.5 matches GPT-5.4 per-token latency in real-world serving, while performing better across nearly every evaluation we measured. It also uses significantly fewer tokens to complete the same Codex tasks, making it more efficient as well as more capable.

Introducing GPT-5.5 A new class of intelligence for real work and powering agents, built to understand complex goals, use tools, check its work, and carry more tasks through to completion. It marks a new way of getting computer work done. Now available in ChatGPT and Codex.

@stevibe tip: use api access for demos

Canceling my GLM Coding Plan subscription. Was running a live demo yesterday setting up a Hermes Agent. Rate-limited after a single message. One. In six hours. On the Pro tier. Switching to Ollama.

@Sophty_ @OpenAIDevs why would you want to use it lol

@OpenAIDevs Can’t build on it when you deprecate it #keep4oAPI

Everyone can just build things.

@Txp_RBI_Xctuxl it sucks

The reason BMW and Mercedes design language has gone from understated class to this cartoonishly gaudy look is because China is one of their biggest markets. It's spiritually 3rd world. It's African dictator tier.

The first simulated image of a black hole, calculated with an IBM 7040 computer using 1960 punch cards and hand-plotted by French astrophysicist Jean-Pierre Luminet in 1978.

cool pi extension from @mitsuhiko! github.com/mitsuhiko/agen…