Rytis

I love coding and to retweet stuff

[Rytis @RytisJM]

Rumors are spreading that with the US gov clearance on fable 5, Anthropic will bring it back on Monday

[Rytis @RytisJM]

If gpt-5.6 was just as good on frontend as Opus 4.8 is it would be no brainer.

Everything codex + Claude design = God level model

[OpenAI @OpenAI]

GPT-5.6 Sol is our most capable model yet for cybersecurity.

It shifts the performance-efficiency frontier for long-horizon security tasks including vulnerability research and exploitation.

[Rytis @RytisJM]

Well I guess big tech got what they wanted. Oh, wait

[Sam Altman @sama]

Good new first: Sol is a smart, efficient, and a significant step forward. It is the same price as GPT-5.5. Also launching in the GPT-5.6 family is Terra, with 5.5-level performance at half the price.

Bad news: at the request of the US government, it is launching today in limited preview instead of the open access launch we were planning on. We are working with the government to get to general availability as fast as we can.

I think it is quite reasonable to roll out models--especially as they reach significant new levels of capability--in this way. It fits with our long-held strategy of iterative deployment. But this isn't quite the process that we think is optimal.

Now we will with the government to attempt to get to a transparent, reliable process for early access, and to ensure that as long as our safeguards work as intended we can release widely. We want to be a reliable, dependable partner that works with all stakeholders, and we also want to live by our mission of benefiting all of humanity. I believe the government shares most of our goals, and that they are overall doing a good job in a very difficult situation.

We will work as quickly as we can to get this model in your hands and we hope you will love it.

[Rytis @RytisJM]

Crazy how china is the "Good guy" in AI now

[Rytis @RytisJM]

One of the biggest concerns is that cutting edge LLMs will be restricted to few people / companies like fable and gpt-5.6 now is and the rest will use significantly worse models.

Best scenario here is LLMs advancing to the point where for most tasks you don't need cutting edge AI anymore.

[Rytis @RytisJM]

US government handpicking who can use LLMs was not on my bingo card for 2026

[Rytis @RytisJM]

@jp54362 For most tasks - yes. Design meanwhile has been lacking behind since day one. I don't think it will change

[Rytis @RytisJM]

Opus 4.8 even at low reasoning levels is very slow compared to GPT-5.5 fast.

Anthropic should really consider giving fast mode to max users. Without fable it's not looking competitive.

[Rytis @RytisJM]

@Revolut Why not acquire PayPal?

[Rytis @RytisJM]

@gmickel Does it work well with the codex MacOS app?

[Rytis @RytisJM]

@levelsio What about version control? For maintenance sounds great, but say you want to build another feature - any simple way to sync?

[Rytis @RytisJM]

@gmickel @cursor_ai With the latest models, how important is repo prompt?

[Rytis @RytisJM]

@gmickel Welcome back! Any plans to integrate Claude code swarm team feature?

[Rytis @RytisJM]

@gmickel @oliverbeavers I found flow next on opencode to be annoying due to the fact that it doesn't finish the epic. Midway it stops and asks "would you like to continue on x.x". Don't have the same problem with the CLI

[Rytis @RytisJM]

@gmickel @convex Why not SQL based db?

[Rytis @RytisJM]

@gmickel I'm interested. Is it 16"?

[Rytis @RytisJM]

@incentive_drift @gmickel Which one works better for you day to day?

[Rytis @RytisJM]

@gmickel @RepoPrompt Would be awesome to try this!

[Rytis @RytisJM]

@gmickel Is Claude code CLI still your go to?

[Rytis @RytisJM]

@gmickel @BuildItWithBen @pvncher @RepoPrompt Is paid version of repo prompt needed to work with flow next?