opinionated 19yo dev i'll build whatever the problem needs.

Poland
Joined November 2025
pcstyle retweeted
Guys! I’m happy to announce I’ve finally gotten over my fear of public speaking!!!!
7
1
1
119
2,594
pcstyle retweeted
Replying to @pcstyle53
Two things that really helped me: 1) Reframing the nerves as excitement! I realized I was invited to speak because people enjoyed my work and were curious about it. The anxiety can also just be excitement about getting to share my curiosity with a room full of people. 2) Staying present. I used to spend the week before a talk worrying about how it would go and replaying past ones where I froze or messed up. Talking with @jackiehluo before my presentation, I realized I’d never actually been present while speaking. I was either rushing to get it over with or reliving bad past talks. Ultimately it’s also about getting the reps in, because a lot of this is easier said than done. But it’s possible to get over it and even enjoy speaking to a crowd :)
1
1
3
42
sooo, my friend - @gdlev_dev just added support for dockers into proxmox? kinda crazy if u ask me. check it out here: github.com/GDLev/dockers-for…
1
1
5
286
ggs
Replying to @rauchg
Looking great! Also started realistic but ended up with a more casual version kickflip.base44.app
4
192
pcstyle retweeted
Replying to @davis7
@davis7 OpenAI is saving your tesla from being wrapped 🙏
I get significantly more cybersecurity risk flags/refusals from OpenAI than Anthropic lately
1
1
9
126
dario. what did u do to fable...
1
8
653
what the actual fuck is this
1
46
muse wonder full review: gemini if its mom hadn’t drunk during pregnancy.
3
72
this guard would rather work for @sama
4
133
`git push`
1
1
8
229
pcstyle retweeted
For the cost of 8 years of $200 Claude subs, you can force your team to use shitty quantized versions of DeepSeek v4 Flash running at 30tps. Incredible deal!
Local AI is good enough for a startup. A 10-person startup built its own 128GB-VRAM AI server for $18,000. One Threadripper box sitting locally with 🎮 4× Radeon AI PRO R9700 32GB 💾 128GB total VRAM 🧠 Qwen3.8 + DeepSeek V4 Flash 👥 16 concurrent AI sessions And they're actually pushing it. Qwen3.8-Flash-Next with 16 simultaneous sessions: 📚 48K context/user ⚡ 538.5 tps aggregate decode At 128K context per user ⚡ 365 tps aggregate decode The GPUs are even undervolted to under 210W each. Link in ALT
817
253
131
10,803
907,696
Switching SMS providers usually means spending a lot of tokens on a rewrite Introducing SMS SDK A simple SDK to send texts through any SMS provider. - has 6 providers (you can bring your own) - very simple syntax - agents can set it up in under 5 mins and it's open source 👀
40
6
4
206
6,864
time to go to sleep ig
5
233
ig meta paid me out too i like the model
MUSE WONDER this is the culmination of over a year of research and infrastructure work in msl since we originally shipped muse spark in april, we have continued to scale our models across pre-training and RL, now reaching the frontier 🏆 see our model's improvements on the artificial analysis index:
2
9
528
is opus 5.5 using `--` instead of `—` in reasoning summaries or is it a @t3dotcodes thing...
1
118
it’s not that far lmao
Let me know when Opus 5.5 can do this
1
10
499