soham

trying to handle errors in TS

[soham @sohamstwt]

@thsottiaux Token consumption is much much higher even for 5.6 terra

[soham @sohamstwt]

okay terra is really token hungry, not like openAI at all

[soham @sohamstwt]

Okay the usage is going down way faster even in 5.5 low

Serious problems

[soham @sohamstwt]

@thsottiaux @thsottiaux can you verify this?

https://t.co/ms0wWTegbf

[soham @sohamstwt]

Okay the usage is going down way faster even in 5.5 low

Serious problems

[soham @sohamstwt]

all good babe, just give us a reset

[Tibo @thsottiaux]

Today, the OpenAI product and engineering team is humming, there is so much to coordinate and keep running across the release of GPT 5.6 Sol/Terra/Luna, ChatGPT Work, the new ChatGPT desktop app, hosted sites and a million other things... together with the new ChatGPT Voice just launched yesterday.

We will not get it all right, give us feedback and we will iterate quickly. Thank you for being an amazing community.

[soham @sohamstwt]

Okay the usage is going down way faster even in 5.5 low

Serious problems

[soham @sohamstwt]

@thsottiaux 🤣

[soham @sohamstwt]

at this point, I'm pretty sure they include gemini as a harassment ritual.

[soham @sohamstwt]

omg

[soham @sohamstwt]

this is a data play. they need good coding session data.

every lab with big bags will do it. anthropic did it months ago. now openAI is doing it with codex.

[zerohedge @zerohedge]

Zuck: “The pricing from some of the other labs is very extreme and has very high margins. We think that there’s a real ability to be able to offer frontier or very high-level intelligence at a much more affordable cost.”

Epic pricing war breaking out among agentic models.

[soham @sohamstwt]

@RhysSullivan here's something I made as a superset of matt's /teach skill.

https://t.co/vuiIMpr9O2

[soham @sohamstwt]

everything comes down to mental models, and exposing yourself to situations where you see the limitations of your assumptions and correcting yourself.

turns out this phd level knowledge factories can do that pretty well.

WHO WOULD HAVE THOUGHT?

[soham @sohamstwt]

I think it will.

I tried @mattpocockuk's /teach skill when making a project with gleam for this first time and it was great.

paraphrasing from @threepointone, agents can compress the feedback loop massively, and that is super helpful.

something I've been working on..

[Ryan Brewer @ryanbrewer]

It’s shocking to me that LLMs didn’t create an educational renaissance. Shouldn’t I be able to learn a language in a month? What did we get wrong?

[soham @sohamstwt]

is discovering that the curriculum design is the most important part for accelerating the rate of learning in a given field.

https://t.co/vuiIMpr9O2

[soham @sohamstwt]

burn your weekly limits, reset coming soon

[Salio @Mr_Salio]

GPT-5.6 Leak

-OpenAI is targeting a July 7 launch
-July 8–9 is the backup window if there are last-minute delays
-GPT-5.6 Sol is already appearing for some Codex users
- GPT 5.5 Sol, Terra, and Luna are all expected to be included with ChatGPT subscriptions
-More generous usage limits than previous GPT models
-Stronger safety systems, but reportedly less restrictive than Fable 5
-GPT-5.6 Sol, Terra, and Luna have all been spotted in the app's code, suggesting rollout preparations are underway

Will GPT-5.6 be the biggest AI release of the year?

[soham @sohamstwt]

Burn your weekly limits before 5.6 drop, they're probably gonna do a reset

[leo 🐾 @synthwavedd]

🚨 SCOOP:

As previously reported, OpenAI plan to launch GPT-5.6 once back in office next week, with a target window of July 7-9, but want it out as early as possible within that window (so July 7th is the most likely date). This is also perfect timing to catch customers coming from Claude who have just lost their access to Fable 5 in plans, and I'm told the 5.6 plan limits will be significantly more generous. More aggressive safeguards are already being rolled out in preparation for the launch too, although they probably won't be as aggressive as Fable's.

DeepMind have also tentatively set a new launch date for Gemini 3.5 Pro of July 17th. Apparently, this extra time has been spent on a new pretrain (they were planning to keep using the ancient 2.5 Pro base, lol), but it remains to be seen whether it's any good. I'm not hopeful. In other news, work is well underway on a new Nano Banana Pro model based on the new 3.5 Pro base, which I expect to be better received and compete well with GPT-Image 1.

[soham @sohamstwt]

The new code rabbit is feature is really good for this.

Semantic ordering of diffs is a fantastic idea.

And it can be done in every other harness too

[Matt Pocock @mattpocockuk]

The "should you read code" debate is dumb because the real decision isn't binary, it's a scale:

1. Reading every line of every diff
2. Scanning every diff, reviewing important lines
3. Ignoring diffs but understanding the 'why' of every PR
4. Spot checking PR's instead of reading every one
5. Ignoring PR's, but doing regular spot checks on the codebase
6. Ignoring the code, but spot checking agent traces to help improve the system
7. Ignoring both the code and the system, let models handle everything

Where are you on the scale?

[soham @sohamstwt]

5.6 can't come fast enough!

[soham @sohamstwt]

t3code is really good at this because i can have a centralized control plane to access and delegate work to all my machines.

what I need right now is a dev environment sync so i can hop between different git snapshots at different machines.

[soham @sohamstwt]

so I've been trying a mini version of this with a hetzner vps.

few pain points so far
- codex sessions doesn't sync across devices. it's such a pain in the ass to tell the agent to sync "/.codex" after few runs.
- ig opencode doesn't too? haven't used it in this process

[Theo - t3.gg @theo]

I have officially moved the majority of my dev work over to Linux. I couldn't be happier. Here's a video about how and why I made the move.

[soham @sohamstwt]

Imagine coding in a language that blocks on I/O