Developer, AI Enginner (in love with AI), thinker
Joined August 2009
- Tweets701
- Following52
- Followers35
- Likes43
I see a lot o people saying there will be an universal income, but never saw a realistic explanation of how it would actually work. Given what I know about economics, I don't believe it work without huge structural changes. Beyond that, it could be more dangerous to depend of private AI companies than it would be if we had conscious AI walking free out there...
Claude is back with todo-list, which is great as we can follow what is being done in a easier way. Codex used to do the same. I think it still does if we ask to.
I'll take a break from @OpenAI codex for a while @thsottiaux . 6.1 sol is great, but it's too slow for me right now. It's taking hours to do not so complicated stuff and that sucks a lot.
codex gpt-6.1-sol today: 30 tps in fast mode.
giving resets to promote this kind of speed is a bad move @OpenAI , @sama , @thsottiaux .
I bet @thsottiaux will tell that there were performance optimizations today. I hope this improvement came to stay!
wow... seems like gpt-6.1-sol had a jump in speed... is that it @thsottiaux , @OpenAI ?
if that's truth, I think it's time to say I love you! hahah
hope not too soon to say that
about new muse model... wait a bit before spending on it's subscription. benchmarks aren't everything.
Having a hard time to burn a weekly usage of @claude code in one day as my reset is coming. Tbh, I am happy with that.
@victornunez
codex cli shows 2x speed and codex app shows 1.5x speed.
which one is true?
why am I feeling @OpenAI will release a model today?
it would be great to get new astra and luna :)
I think @OpenAI has good models, but right now they're behind @claudeai .
Right now I see two things that might be happening:
- OpenAI may actually make money with the subscriptions given the latest moves I've seen them doing.
- Anthropic and @elonmusk might be about to go in a closer partneship (I mean really closer)
Let's see what future brings. Competition is great!
I'll tell my experience with codex gpt-sol-6.1 so far: I's doing a great work, but sometimes it's very slow. I am very pleased with usage limits, but I am on 20x yet, so if I think about it - no. Usage is not as good as I'd expect as I would be able to only do HALF I do right now. If I could keep my usage for the same price right now it would be a great deal.
Something's odd about PRO 20x subscription. As far as I know it would stick as 20x until the end of the month, but right now it's burning too fast.
@OpenAI , did something change? It was harder to spend usage before, but now it's kinda easy...
Dots from @OpenAI is just unusable to control and command projects. every stupid decision it has to take it prompts you and even when you allow it it keeps asking permission because it's unable to allow it itself. annoying. It's promising, but too much constrained. Unusable.
lol
just tried to call my @OpenAI dot and it took very long to attend.
dot was budy, funny haha