
building https://teapawt.net/t.co/GDwiLa7WnI ex poker-pro, @accel
Joined April 2021
- Tweets975
- Following657
- Followers2.1K
- Likes3.3K
true story: a friend bought bitcoin ~10 years ago on some sketchy site that’s long dead. he did not remember which site, which wallet, or where the keys went.
so he told codex to find his bitcoin. gave some vague details.
it dug up an old file buried on his disk with the purchase details.
he’s given his agent his whole digital life. so have i. yes, more attack surface. but people in my friend group are getting very ai-assistant-pilled.
2
14
1,209
India is becoming the world’s exporter of RL environments.
Edtechs sell environments their students build. Quick commerce sells warehouse robotics data and buying data. Home services sells robotics data. Healthcare companies selling health data.
We may have lost Infosys and TCS to AI. But until labs reach RSI, data is India’s next billion-dollar export.
anthropic could double the price of their max plan tomorrow and nobody would leave.
will people crib? yes.
will they churn? highly doubt it.
that’s the IPO thesis.
5
338
what crypto bros are worried about: memecoin scammers being exposed. democrats coming after crypto companies. devs leaving for ai
what they should really worry about: the guy who bought the bitcoin top 5 years ago is sitting on gains lower than treasury yield while everyday worrying about when he would get hacked
5
452
Nikhil R retweeted
This is so cool to see come to life! Had so much fun working on the branding and visual identity for this.
Super excited for the product and the absolutely cracked team behind it.
I’m sure @manasjsaloi is going to kill it. Can’t wait to see where this goes!
Come trade stocks on Agara
Up or down, 5 minutes to daily
0% fees. Self-custodial. Settled on-chain
app.agara.xyz
Everybody starts somewhere
Super juicy BTC funding arb available in 7 figs on @HyperliquidX
Almost 400% annualized to short, while CEXs are flat for an easy hedge.
Also a record volume day for BTC on the platform!
1
1
356
Nikhil R retweeted
At least $75K was lost by copytrading bots today by an exploit from malicious actors
This is how it works and how we protect you from this on betmoar 👇
nobody’s paying me to say this, i’m just biased because we built it 😝
agara has 0% fees on short-term crypto and stock markets and on everything else. and you get markets kalshi and polymarket don’t even list, like your own country’s inflation and macro markets.
try it out at app.agara.xyz
one of the biggest bets i’ve ever taken was kamala to win in 2024. most of my trading till then was stocks and crypto, this was my first real position on a macro event
didn’t even know how to hedge it properly. but in my head it was simple. harris wins, i make a lot of money and my crypto gets nuked. harris loses, i lose the bet and my bags pump
she lost, bags pumped. wonder how many people are running similiar trades for the midterms
6
544
genuinely wondering what @Kalshi and @Polymarket employees are rooting for tonight
dems win like the markets are saying and you get validated, but now you probably get more people in congress who want to regulate you. or the markets are wrong and you keep the friendlier regulators but look stupid
interesting day for them
5
464
i know i am in a bubble but it’s kinda funny that the way a lot of us follow elections now isn’t tv news or exit polls. it’s refreshing kalshi and polymarket
only true for the US right now but it’ll be true everywhere soon. people can keep saying prediction markets are just sports and crypto, but for a lot of people globally it is just: news
Nikhil R retweeted
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! 🧵
Blog: rikiyat.github.io/blog/weigh…
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math.
rikiyat.github.io/blog/weight-sp…
github.com/RikiyaT/lm_und…
huggingface.co/Qwen/Qwen3.5-0…
huggingface.co/RikiyaT/Qwen3.…





















