Theo - t3.gg

59.6K posts

Theo - t3.gg banner
Theo - t3.gg

Theo - t3.gg

@theo

Full time CEO @t3dotchat. Part time YouTuber, investor, and developer

San Francisco, CA Katılım Ekim 2016
4K Takip Edilen302.1K Takipçiler
Sabitlenmiş Tweet
Theo - t3.gg
Theo - t3.gg@theo·
So I've been using a new terminal...
English
73
24
807
135.9K
Theo - t3.gg
Theo - t3.gg@theo·
Big T3 Code drop incoming. Should I go live for the release?
English
6
0
20
510
Dare
Dare@zenbyteio·
@theo Did you get the ok? Or should we just go rogue? 😂
English
1
0
1
447
Theo - t3.gg
Theo - t3.gg@theo·
@thdxr They need to steal another 150,000 prompts from Claude first
English
16
0
191
6.1K
dax
dax@thdxr·
where the heck is deepseek v4
English
27
6
291
14.4K
John
John@jrysana·
@theo No, not if it's set up properly, the larger model still runs all the exact same calculations just with more parallelism and lower amortized weight loading overhead
English
1
0
4
362
Theo - t3.gg
Theo - t3.gg@theo·
Can someone smarter than me explain how these “turbo mode” things work on like an implementation level? Are they using better GPUs? More GPUs? Reserved allocation during generation?
English
53
1
306
44.5K
Eric Parker
Eric Parker@EricParker·
@theo Batching choices. Fast tps for a user means less system wide.
Eric Parker tweet media
English
1
1
62
3.4K
Theo - t3.gg
Theo - t3.gg@theo·
@alhasaniq None of these labs are using Cerebras for their flagship models or flagship “fast mode” offerings
English
1
0
7
2.3K
Not Not Alhasan Al-Nasiry
@theo Without degrading quality: - Cerebras instead of nvidia (muuuuch larger die) - Less interactivity (less concurrent users per gpu cluster) With degradation: - Quantization - Expert pruning
English
4
0
3
5.4K
Theo - t3.gg
Theo - t3.gg@theo·
I thought I was dumb for not knowing the answer to this. Seeing how dumb ~50% of the replies are, I feel less bad now. Thank you all 🫡
English
1
0
46
4.9K
Theo - t3.gg
Theo - t3.gg@theo·
@jrysana I thought speculative decode would affect the output quality in some meaningful way, does it not?
English
4
0
3
2.2K
John
John@jrysana·
@theo Better GPUs (Blackwell) Speculative decoding Lower batch size
English
3
0
11
4K
Theo - t3.gg
Theo - t3.gg@theo·
@Croczillak Am I allowed to also pay API prices when I go over my usage limits?
English
2
0
4
1.2K
Mike
Mike@Croczillak·
@theo If you were restricted to one, Cursor sub or GPT sub?
English
1
0
0
1.3K
Theo - t3.gg
Theo - t3.gg@theo·
Cursor just dropped a T3 Code clone
am.will@LLMJunky

Nice. @cursor_ai just dropped their new "Glass" alpha, and they're leaning heavily into the simplified coding GUI trend that's been blowing up lately. First impressions are really positive. And just look at how insanely fast Composer 2 is. First impressions? Drop yours 👇

English
35
7
424
43K
vladkens
vladkens@uid127·
@theo It's so sad they just didn't buy your project.
English
1
0
1
1.6K
Theo - t3.gg
Theo - t3.gg@theo·
Fwiw, I think everyone is going to do this now. It’s a much better UI for the “new way of coding”. Cursor’s main UI has been in a rough state for like a year now. I really like that they are resetting. Overdue tbh. Also really like that they support ACP, might have T3 Code support in near future? 👀👀 (Composer 2 is really good, don’t sleep on Cursor) (Also insert obligatory *I’m an early investor with early access* bias accountability)
English
16
0
183
9.7K
Theo - t3.gg
Theo - t3.gg@theo·
@alexinexxx I know cachy is the over hyped option now but it’s very very nice
English
2
0
25
2.4K
wearifulpoet
wearifulpoet@wearifulpoet·
@theo I’ve always wondered which is a bigger factor: the influence on your thoughts or the confirmation bias of noticing ads that match your thoughts?
English
1
0
80
39.8K
Theo - t3.gg
Theo - t3.gg@theo·
@AprilNEA I’m guessing a lot of the deploy part is just from bun tbh
English
0
0
24
2.2K
AprilNEA
AprilNEA@AprilNEA·
🧵 I just reverse-engineered the binaries inside Claude Code's Firecracker MicroVM and found something wild: Anthropic is building their own PaaS platform called "Antspace" (Ants + Space). It's a full deployment pipeline — hidden in plain sight inside the environment-runner binary. Here's what I found 👇
AprilNEA tweet media
English
62
186
1.5K
216.4K
Theo - t3.gg retweetledi
NeetCode
NeetCode@neetcode1·
I just met the creator of react
NeetCode tweet media
English
90
30
4.1K
137.8K