Been interning at @ThineAI for a few months now, and reading this interview made me reflect a little.
I remember my first conversation with Pratyush and thinking, "If I get this opportunity, I'm def taking it."
A few months later, I can confidently say it was one of the best career decisions I've made. Great founders attract great people, and I've been lucky to learn from both every day.
Forever grateful ❤️
marketing team said "just tweet something genuine."
so genuinely this cafe video shows thine doing exactly what we designed the context engine to do.
and i'm genuinely being forced to tweet this. both things are true
Your brain isn't built to be the entire team. @ThineAI is.
It holds the context of your business. Orders, negotiations, emails, opportunities.
No long prompts. No complicated setup.
It just pays attention. Follow @thineai & Join the waitlist!
spent months making sure thine's context engine doesn't miss a beat. meanwhile i missed my own dentist appointment twice this month. building for thine is humbling
tag @thineai for a private beta invite
Your brain wasn't built to hold everything. So we built @ThineAI
It runs in the background. Captures the context of your work and life. And gives it back when you need it.
Comment thine and we'll DM you a beta invite.
BREAKING: Jensen Huang boarded Air Force One at the last moment.
Elon still under investigation. Trump taking charge.
The China Visit, only on TechBrunch.
A lot of “bad memory” behavior is really a timeline problem.
The system may have the right facts stored, but it cannot tell which fact is still active, which one has expired, and which one represents future intent.
That is much harder than simple recall.
We are solving this with our latest Temporal Reasoning algorithm release.
starting to think now that every agent should have just 2 tools. search and execute. we _want_ agents to have access to 100s, if not 1000s of capabilities, that can contextually change during their lifetimes, even per message. saying stiff like "just use bash" doesn't encompass 3rd party apis, and you don't want to keep switching up the base prompt all the time. you gotta generalise that. I also guess search has to be semantic, so probably something with a vector db type thing. does it run on every message? probably...