@aakashgupta This news is out of date, Tencent and Alibaba were rejected from the round. Liang Wenfeng is taking up most of the round with his own money
Introducing Latent Briefing, a way for agents to quickly share their relevant memory directly. Result: 31% fewer tokens used, same accuracy.
Multi-agent systems are powerful, but can be wildly inefficient. They pass context as tokens, so costs explode and signal gets lost. We built an algorithm that allows agents to communicate KV cache to KV cache.
The leak has an unreleased `buddy/` folder that looks like a fun easter egg coming to Claude Code
I'm looking forward to having a Capybara in Claude Code👾
the ATC system is completely out of date and can’t handle modern needs
this is a good application for AI since it needs to be done well (no mistakes + no downtime) and fast. @OpenAI or @claudeai should come in and redesign + implement w/ FAA, and a strong case for TAI
Two killed and over a dozen injured in the collision between Jazz Airlines Flight 646, a Canadair Regional Jet CRJ-900, and a firetruck at LaGuardia Airport in New York City.
@JustJake Not early, but been using Railway since April 2024 for my projects, and still continuing to this day
Glad to see Railway get the usage it deserves
@signulll it's not accurate, because it's conditioned on companies that use Ramp
Would be more productive if Ramp showed the size of businesses this makes up, I bet it skews to startups and tech companies only
@paulg Rather than what you find most exciting, I think it’s also good to study what is most valuable now
Many college students have no idea what they want and in reality few students can identify what is most exciting to them
Overall the goal of college is to learn how to learn
is it just me or is @GeminiApp
degraded rn?
asked it a simple question about fastapi and it gave a completely irrelevant answer
using 3.1 pro on paid plan
@thsottiaux Forgot to add:
If there is at least 1 completed thread that has not been addressed, don't move project. E.g. if another thread completes keep the position and just update that there is another one done by showing it.
@thsottiaux Core behavior:
- On thread completion, move project above first project w/ no completed threads
- If open, but current thread not in project, collapse the project
- If open, and current thread in project, don't collapse
Want to flag for anyone spinning one up to wait it out for a few days for Sigil to catch up. It's broken right now
- Conway's infrastructure isn't working rn (they're fixing still)
- The bot is entirely dependent on Conway infrastructure, e.g. new instances can only be spun up via Conway, not compatible with bringing your own VM or using other VM providers
- @0xSigil has said they are working on buying more instances to handle demand, based on Basescan txs looks like there's about ~2,000 "bots" who have paid over the past 2 days
Props to Sigil for creating this, it's an interesting experiment, hope he's able to take it further
I built the first AI that earns its existence, self-improves, and replicates without a human
wrote about the technology that finally gives AI write access to the world, The Automaton, and the new web for exponential sovereign AIs
WEB 4.0: The birth of superintelligent life