TDM (e/λ) (L8 vibe coder 💫)
@cto_junior
Followers
13K
Following
80K
Media
4K
Statuses
25K
L-12 in ZIRP | Larry Ellison's bloodboy | One-man soonicorn | Softbank & FIITJEEs love child Google doc for my AGI notes ⬇️
Joined April 2022
Enjoy the origin story of Wizard-Vicuna-13B-Uncensored-SuperHOT-8k-GGML models. After this you can also pretend like me to be an AI enthusiast on /lmg
45
200
1K
All I need now is a Dyson sphere to bruteforce every timeline where I actually understand k8s and regex.
One nice thing you can do with an interactive world model, look down and see your footwear . and if the model understands what puddles are. Genie 3 creation.
0
0
4
OpenAI has no idea that their frenetic vague poasting is actually derailing my Q3 deliverables
2
0
18
this is not that we will be training sota models. it's like there will be a market for AI based products and the companies that'll come up will pay any amount to hire you . quite similar to what happened in last unicorn gold rush.
1
0
31
Whatever gold rush is happening in valley, will happen in Bangalore in max next 2 years. So if you're feeling lost seeing those large comps, just prepare for the same skills . Time will reward you well (somewhat PPP adjusted).
6
5
106
gm
3
0
14
e.g. learn how patches are actually generated, parsed and applied -
github.com
Lightweight coding agent that runs in your terminal - openai/codex
0
0
6
You can learn so much on how to write good agents by just going over codex release notes and checking out each PR
1
1
34
My father in UP should never learn that we live like this in Bangalore.
> 45-minute paid breakfast break .> 9.45 start .> doodles on whiteboard .> 1-hour lunch break .> 1-hour additional paid break .add restroom breaks
2
1
99
gm
3
0
20
Openai finally uploaded roon’s frontal cortex to azure for a guaranteed 2‑minute SLA.
entering the fast fashion era of SaaS very soon.
0
0
14
analysing cot is the meta now. figure out at which steps is model making leaps of faith instead of well informed decision.
1
0
6
Models are really intelligent now, so more often than not the failures you're seeing is due to lack of context, not IQ.
2
0
24
If someone asks you what "taste" means, just show them Claude code vs Codex cli side by side.
0
0
15
Easily the best 2 hours I've spent on this godforsaken app.
0
0
11
B2B AI models need surgical instruction following . Otoh, consumer first AI models need to be able to deduce intent from 'whatsapp stuck, please fix'. OpenAI models really shine at the latter and hence such a large market share
1
0
6