tnmy

AI, Urban Planning, Farming
jack of all master of some :)

[Reads with Ravi @readswithravi]

I’m in love with this sentence:

“Every pattern in your life repeats until you learn the lesson. The moment you choose differently, the loop ends and growth begins.”

[tnmy @tnmy909]

@RaniChatterjeee @grok whats that book? Animal farm?

[tnmy @tnmy909]

Why are there so many VideoGen and Editing startups in @ycombinator
All of them are Higgsfield alternative but specialized, and almost all of them are credit based like higgsfield is.

A total of 21 from YC25 and YC26. List below :

YC25 :-
• Pickle
• Mosaic
• Narrative
• Knowlify
• Koyal
• Dollyglot
• Bluma
• Golpo
• Stewdio
• OnDeck AI
• cocreate

YC26 :-
• Cardboard
• Wideframe
• Martini
• Flick
• Grocalo
• Palette
• Vyra
• Absurd
• Voyager
• Diffusion Studio

Why do I feel more than half of them are going to die soon🥲
This looks too saturated, and YC is just betting on all of them for what?

[tnmy @tnmy909]

@nsitharaman @AmitShah @grok how many days PM and HM attended parliament in Monsoon session? Don't lie to me I watch LSTV everyday 😂

[tnmy @tnmy909]

@shydev69 Aaj toh mangalvaar hai, hanuman ji ne bhi saath chhod diya😭☠️

[tnmy @tnmy909]

16gb ddr4 from pre-ai era😓

[Mia @MiaAI_lab]

Memory is the new GPU ‼️

Apple's new CEO: "What we're seeing with memory pricing right now is unprecedented."

To put it in context, DRAM contract prices rose 93-98% in Q1 alone, 40% in Q3, and expected to rise another 20% in Q4.

How much RAM you got at home???

[IRIS C2 @C2IRIS]

Asked one of our developers how he uses his AI agents and he just sent me this:

[tnmy @tnmy909]

@abhishekcode42 @grok wtf???!??

[tnmy @tnmy909]

@LH @ZedLang @grok wooahhh new language sounds cool.. Can you brief on how it compares to likes of Bend or Rust or something other new language?

[James Whittington @jcrwhittington]

Today we are happy to share one of our first open research releases: Latent Sparse Attention (LSA), a step towards subquadratic attention that matches accuracy of full attention at ultra-long context.

Sparse attention attends to only a few tokens per query, but it has to find the right ones. At 10M tokens, that search takes 98% of the attention layer's time. LSA makes sparse attention up to 22.7× faster and full attention 72.4x faster.

With @kubicek_ales @caciolai @PengLiangzu @SukritSumant @Ben_Hoov @FraserGreenlee @RMBattleday @radim_urban @thoefler @jcrwhittington

[tnmy @tnmy909]

@grok Cool stuff, could be paired with AI, home assistants etc for a smart home experience.

[tnmy @tnmy909]

@grok sometime ago I saw a creator making small metal things (size of shirt buttons) which produce a specific sound based on their shape (the way they are cut), and each button has a unique sound which shows up on spectrogram. It can be used for detection, So one can put a button on door and then moment door opens it will show unique "wave" in the software, or another button on fridge which has its unique sound and so on...

Can you find that thing/creator for me pls.

[tnmy @tnmy909]

@0xSero Don't forget water when fasting. Stay hydrated, stay safe <3

[Lessons @onlypreneurs]

In a weird way, everybody is rooting for you. When you go to a job interview, the interviewer wants you to be amazing so they don’t have to look anymore. when you go on a date, the person wants to have an amazing date with you. If you make a piece of content, people want to see amazing pieces of content and they have no problem if it’s yours.

the instant you get into the mindset that people want you to do well, your entire world gets better. And you’ll be shocked to know this pattern keeps repeating in different areas of life.

[tnmy @tnmy909]

@thsottiaux When will you announce 2x limits for plus users ; business plus seats already got it!

[tnmy @tnmy909]

I knew Yann LeCun was onto something when he said "world models" few months back, and I resonated with some of his philosphy about AI and intelligence;

--- but does it really matter if you're "early" to something if you don't actually materialize that. i am spotting a pattern in my life. I was early to claude, sonnet 3.5 being the first AI i used. I was one of those guys in college who refused AI for research for a year and i produced better results than my peers who used GPT.
I was "early" to crypto (relative to people around me), saved up money and brought Ethereum back in 2020 and held it for few years.
I was early to online casino (stake), moved that ETH i saved up to stake and made money from it and exited. My friends got into it after me, lost their money (they still sometimes break my balls over it)
Then I got into agentic AI (opus 4.5), made my first project and my own website (which i havent marketed anywhere or updated)..
And before any of that I started using linux just for the fun of it, still daily driving for 4 years now.

The point being, I was early, I saw the potential and future, I tinkered and played and enjoyed, but I never got serious about it, never committed. Thats a problem. But hey atleast I'm aware of it, so thats good right??
No its not. I have abused my awareness for too long... I gotta fix it, I gotta fix myself, I gotta lock the fuck in or I'm ngmi.
Its like god is telling me hey you can do this or that and I'm like nah, i won't do it just because you told me to. But thats just my adhd i guess...
This is a public confession, maybe....

"The way you do anything, is how you do everything."
I MUST MATERIALIZE MY IDEAS OR THEY ARE WORTHLESS.

[tnmy @tnmy909]

You've been carrying the guilt of quitting something you never began.

[tnmy @tnmy909]

rules that are too strict tend to get broken all at once.

[tnmy @tnmy909]

Biology guys in Tech is refreshing to see.
I recently found out that Dario has a PhD in Biophysics.

Man x Machine is the future and we need Multi - domain research institutions across the world, that is the future!

[Arnesh @arnesh_11]

we’ve built a neurosymbolic AI that can chat like an LLM.

it just scored 28.4% on ARC-AGI-3 (on the kaggle leaderboard) for $19

for context:
GPT-6.1 Sol (High): 26.72%
GPT-6 Astra (Low): 17.45%
GPT-5.6 Sol (Max): 7.78%
Claude Sonnet 5.5 (XHigh): 24.31%

unlike other ARC-AGI Competition submissions which use an LLM and harness, our AI is a beta version of a symbolic ai + activation steered RL-based transformer (in the works).

we have a main inference harness and a browser-use and computer-use harness ready.

now we just need to find out how to get a 96GB VRAM GPU server to make it publicly accessible 🥀

[tnmy @tnmy909]

@Bhavyaztwt Fight me!
https://t.co/9eSX4W73nE

[tnmy @tnmy909]

These are the best distros!

(Quattro, Nobara, Mint)