Developer, AI Enginner (in love with AI), thinker
Joined August 2009
- Tweets955
- Following56
- Followers44
- Likes75
Claude code's sub sessions functionality is amazing. If you want a manager to control multiple work, it does a pretty decent job, awareness of each other session is just great and visual organization is just amazing.
1
14
I just love AI. It's probably taking my job away (I am a developer) but I love it so much, because it makes my creativity fly high!
2
2
29
interesting take: the execution time for gpt-6.1-sol in @ArtificialAnlys considers it run at 54/tok/s. In reality it runs at about ~30 tok/s which is 55% of that measured speed. That means the the execution time we see in artificial analysis are not real. The real speed is twice as slow as what they're showing.
1
2
27
Idk why nobody is talking about it, but I really feel my ultra subscription has dropped a lot in usage.
Usually my subscription used to endure like 2-3 days. Now it did endure a bit more than one. Curiously right after becoming a 10x sub. @lydiahallie , did something change in the 200 usd subscription usage?
13
That's just incredible how waiting for a new model gets you anxious. Am I addicted because of that?
I think I am. This whole week one of the first things I thought about was about getting a smarter, faster and cheaper model. Just too frustrated about it.
I think that @Ananth7e, @kimmonismus and @TokenGremlin feels the same way according to what I've seen. ☹️
1
2
50
I hope that when new models drop (hopefuly today) we get a reset along with it @thsottiaux @ajambrosino :)
17
$500 subscription enduring 1 day sucks @thsottiaux , @victornunez and @reach_vb .
I did reset yesterday and right now I am at almos 10% of usage remaining. And the work I've been doing is the same I was doing since saturday at night.
I think it could be related to the issue I mentioned: selecting ultrafast in one session is making other sessions to run in ultrafast too, which is a big problem.
Already reported it and had no response up to now.
26
jeez.
Astra is a great model... but it gets distracted very easily.
if you add a follow up or a hook runs, it just forgets what it was doing and that's pretty annoying @thsottiaux . it would be great if it could be fixed.
19
I wonder of I should give a try to @Meta muse code.
Does anybody here have tried it?
What you think. Does it work to give a try?
When it was released I tried their plan code, but it was really bad.
51
This is great! Congratulations @neuralink ... this is def one of the most exciting kind of science!
Our engineers recently visited one of our speech restoration study participants for a 72-hour hackathon.
The goal: rapidly prototype as many features as possible to enhance the experience of using Neuralink to communicate with thought.
Here are some of the things we built 🧵
16
ultrafast right now:
@OpenAI for god sake... this needs to be fixed, or if speed is slower than minimum expected, user should be charged accordingly. This is just too bad.
51
I see @ArtificialAnlys say's 6.1 sol speed is about 79 tok/s. Where is it? I don't see such speed anywhere.
1
21
This is a big win for users!
Day 25/
We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium, effective today.
We heard your kind feedback and won’t be reducing usage limits for Pro 200, so your Work and Codex usage will stay the same. We’re also going to continue offering extended limits for dots for now, and usage of your dot will still continue to be separate from your default usage.
And we'll of course also keep shipping better models and other efficiency improvements that increase your effective usage over time.
12
in the beginning terra 6.1 was about 150 tok/s
now it's about 120 tok/s in fast mode avg
wtf @thsottiaux
14
is deepswe @datacurve dead?
they haven't updated it, and artificial analysis stopped using it as well.
just curious.
know something about it @grok ?
1
22