laxman

3K posts

laxman banner
laxman

laxman

@llmluthor

20 | It is our moral obligation to make an impact in the world proportionate to our potential | shippin @vektorimemory prev: cognitivelab, bhindiai(upsurgelabs)

Katılım Eylül 2022
365 Takip Edilen202 Takipçiler
Sabitlenmiş Tweet
laxman
laxman@llmluthor·
We built an AI that remembers like a brain, not a database. Vektori builds a living graph of every conversation you've ever had. Fresh concepts glow bright. Old ones fade, unless you reinforce them. 20+ daily users trusting us with their entire digital memory. No venture backing. Just pure engineering spite :)
English
6
4
18
1.8K
laxman
laxman@llmluthor·
CoT changed the world
English
0
0
0
10
laxman
laxman@llmluthor·
@aaron_epstein Yess man, makes so much sense. First go so deep into people's problems and early on just vibing with them to understand their problem, then go pitch your product and even after that if the vitamin you prescribed is not right, change the product Or customer segment
English
0
0
0
15
laxman retweetledi
Aaron Epstein
Aaron Epstein@aaron_epstein·
Think of startup sales as a doctor/patient relationship. If you just walked into the doctor's office and they said "here take this prescription", you wouldn't trust it at all. They don't even know what problems you have yet! Sales is the same. You are the doctor/expert that has to ask lots of questions to understand your customer's problems before you earn the right to show how your product can help. (shoutout @ryanRneu for teaching this to many founders)
English
32
28
308
22.9K
laxman
laxman@llmluthor·
Tho I also like steve jobs version, where the whole point of building career is broken, as steve jobs says: "Do something you enjoy so much that you can't stop thinking about, and you only sleep to wake up to do it all over again"
English
0
0
2
22
laxman
laxman@llmluthor·
Career wise, the comfortable choice is very often the wrong choice
English
1
0
2
30
Indilingo
Indilingo@indilingo·
Excited to share that Indilingo has been awarded $2,500 in API credits from @OpenAI 🎉 As a bootstrapped startup, support like this helps us experiment faster, ship better products, and continue building AI-powered learning experiences. Grateful to OpenAI for the support ❤️
Indilingo tweet media
English
11
5
34
1.3K
laxman
laxman@llmluthor·
@sama Or you scammed apple?
English
0
0
1
21
Sam Altman
Sam Altman@sama·
there are a lot of benchmarks that suggest 5.6 sol is the best model in the world right now, but the most reliable way to tell is that elon is obsessed with me again
English
4K
2.9K
62.8K
5.7M
laxman
laxman@llmluthor·
@sama homeboy did you steal from apple?
English
0
0
0
21
laxman
laxman@llmluthor·
everyone in memory space is over engineering retrieval, graph layers, rerankers, fancier RAG. we did too. turns out that's not the bottleneck. so we trained a small model to decide what memory keeps, drops, resurfaces. come see it live- Agentic Summit BLR x Magicball, July 22.
English
1
0
5
255
laxman
laxman@llmluthor·
Winners never quit, and quitters never win.
English
0
0
1
58
laxman
laxman@llmluthor·
@1owroller or maybe, she's avoidant? do you believe in attachment theories?
English
0
0
2
714
laxman
laxman@llmluthor·
too much vibing
laxman tweet media
English
1
0
4
102
raveeshu pahuja
raveeshu pahuja@PahujaRaveeshu·
Today's agents forget what matters — we're fixing that in latent space. Grateful to the community that stress-tested the idea into a company
South Park Commons India@spc_india

SPC members often find themselves in a bit of a mess. The messy ‘-1’. @PahujaRaveeshu was there, after moving back to India from USA, to explore curiosities. He went looking for a ML paper reading group, couldn't find one, and decided to start one here. That was then.

English
1
0
1
78
laxman
laxman@llmluthor·
new look, same problem. say your agent needs a skill so you spend a weekend writing it then testing it and then forgetting it exists. built agentnet so you don't have to do that anymore. 100k skills, one install.
English
3
2
12
1.6K
laxman
laxman@llmluthor·
pip install agentnet-cli
English
0
0
2
93