Adi retweeted
If he had asked even one tech person on X before posting this, he could’ve saved himself
What are you even trying to sell man?
This isn’t YouTube. People here can see through this crap
Adi retweeted
Indian company interview assignment:
Our company cost of LLM is drastically increased
Build 105B llm and benchmark against opus5.5
1 week timeline, internet allowed
Adi retweeted
yesterday turned out way better than we expected. crossed 100k views and 12k site visits with people from 129 countries checking out Engage and US driving the most traffic
@devsvn2 and I couldn't have asked for a better start. huge credit to the team behind this
Adi retweeted
We at tryengage.si are trying to save students' time and get hired fast.
Adi retweeted
Sarvam is as distracted as me.
Just that it is funded and I am not!
Adi retweeted
Anthropic’s second door for people who don’t get in
tryengage.si
Applications are now open for the Anthropic Residency, a paid, fixed-term research program for academics, technical founders, and staff-level engineers who want to make AI safe and beneficial for everyone. No ML background is required.
Apply by November 22: alignment.anthropic.com/2026…
Adi retweeted
A product can have incredible features and still fail.
Because customers don't buy features.
They buy a better version of their current situation.
That's a distinction more founders should think about.
Adi retweeted
how do big models fit on smaller hardware?
quantization: store each weight with fewer bits, 4-bit instead of 16-bit cuts weight storage about 4x.
whether the quality holds up depends on the method and the model, which is why new quantization papers keep landing.
Day 28/ tibo's game has landed
i spent $5.99 on this domain, but i had a blast making it with grok bot and cursor! i hope you enjoy it and find all the easter eggs 🤍
if i did it right you should be able to play it inside of X too
tibo.dog
Adi retweeted
there are tokens in gpt's vocabulary that were never in its training data, because the tokenizer and the model were trained on different corpora.
"SolidGoldMagikarp" is a real token. ask the model about it and watch it hallucinate on command.
WisprFlow marketing hire #2 has to work extra hard now.
We’ve raised $1.3M from YC & Afore Capital so you never have to type or speak again.
Typing is slow. Talking into your computer is awkward. Subvocal fixes both
Our glasses give you the speed of speech without you making a sound or touching a keyboard.
It tracks your internal speech muscles with non-contact sensing and decodes silent, almost invisible mouthing into text at more than 200 words/min.
Pre-orders are now live at:
subvocal.co/preorder?utm_sou…
Adi retweeted
If I am applying to a YC startup, here's what I'd do (Got this from someone who actually got hired at one )
> Find the founders on X and LinkedIn, see what they've shipped recently.
> Spend an hour or two exploring their product. Understand what they're building and what could be better
> Build something for them. A landing page redesign, a small feature, or even a bug fix. Show them what you can do.
> Reach out with a personal message. "Hey, checked out your product and noticed X. Built something that might help. Would love your thoughts."
> Avoid generic cold DMs. Be specific. Show that you've actually put in some effort.
> Follow up. Prepare well for interviews. Show that you're serious about joining their team.
Stop sending YC founders bulk automated emails for hiring. Take time to write a personalized email explaining why you want to work with them. Spend just 5 minutes with their product. That’s literally it. By spamming you are just gambling with your chances to get hired. They will probably stop looking into emails it this trend continues.
Adi retweeted
Periodically it should be Barium/Berkelium/Beryllium/Boron/Bromine after Argon
But they don’t have compute to cover them all
I did say we'd be releasing our models periodically...introducing Gemini 4 Carbon, our next frontier model.
Following Argon's release just two weeks ago, Carbon shows the pace at which we’re pushing the frontier, it ranks #1 on the recently updated v5.0 of @ArtificialAnlys Intelligence Index at 59. It also leads across key cybersecurity evaluations scoring 88.2% on SEC-bench Pro for finding vulnerabilities, 54.8% on ExploitGym for exploitation, and 71.0% on CWE-bench for patching.
We’re seeing significant improvements in long-horizon tasks even compared to Argon, and we’ve strengthened frontier safeguards.
As we did with Argon, we’re going to roll it out gradually + responsibly. Some trusted cyber defenders will get access starting this week. We'll be expanding it to more partners through our Fairwind Program, ahead of wider availability soon, so hang tight!
Adi retweeted
a lot of people want to learn rag but don't know what to actually read.
7 papers i'd recommend, in order:
> dense passage retrieval (dpr)
> retrieval-augmented generation (rag)
> hyde
> lost in the middle
> self-rag
> crag
> graphrag (from local to global)
by the end you'll know most of what production rag systems are built on.
A month ago there was a launch video from tavas and you could say that it was AI. But this ?? You can’t possibly say it’s AI without reading the caption. Just insane.
Very excited to launch Avatar Arena! 🤖
Video is the highest-trust medium for human communication, and when it works, it works better than anything else. And, AI models have just started to come around to supporting high-quality, studio-grade video generation.
If you're someone with use cases for AI video - like product demos, ad hooks, founder updates, or podcast promos - we'd love for you to try us out. You can generate avatar vids by recording a 15-second clip of yourself or just going with a stock avatar. Two AIs will give you their best shot of bringing your vision to life and you choose one!
Avatar Arena currently supports models from - @HeyGen, @synthesiaIO, @veedstudio, @Creatify_AI , Captions, @Jogg_ai, @tavus and @runwayml.
All the votes over the next few weeks go into a leaderboard, which we'll publish soon.
Try it @ arena.newmeasure.ai :)
(Could you tell the vid below is AI? 👀)
Adi retweeted
I wanted to hire a video editor
But for a benifit of doubt, i gave opus 5.5 a try
And never going to look back🔥🔥
This is AMAZING 😭😭🔥























