spolu

9 ms·

1.
Code Contracts: Structured specifications colocated with code(code-contracts.cc)
4 points by spolu 6d ago|0 comments
2.
Hive: Engineering at the Speed of AI(dust.tt)
2 points by spolu 8mo ago|0 comments
3.
Build vs. Run(dust.tt)
1 points by spolu 8mo ago|0 comments
4.
Infrastructure behind Dust deep-dive agent(blog.dust.tt)
3 points by spolu 11mo ago|0 comments
5.
Network of agents collaborating through a publication/review system(github.com)
2 points by spolu 1y ago|0 comments
6.
Show HN: Dead simple transcription command-line(github.com)
2 points by spolu 1y ago|0 comments
7.
Thoughts on the Future of Coding(twitter.com)
1 points by spolu 1y ago|0 comments
8.
Cybersecurity and LLMs(twitter.com)
1 points by spolu 1y ago|0 comments
9.
The Visions of Neil Mehta(joincolossus.com)
2 points by spolu 1y ago|0 comments
10.
MCP and the emerging enterprise AI OS layer(blog.dust.tt)
2 points by spolu 1y ago|0 comments
11.
by spolu 2y ago
Yes you are perfectly right. Our product pushes users to be selective on the tables they give access to a given agent for a given use-case :+1: The tricky part is correctly supporting multiple systems which each have their own specificity.
12.
by spolu 2y ago
Hi, you are right that things can go sideways fast. In practice, the data that the typical employee needs is also quite simple. So there is definitely a very nice fit for this kind of product with a large number of use-case that we do see p
13.
A fair comparison of modern models' reasoning capabilities(blog.dust.tt)
1 points by spolu 3y ago|0 comments
14.
Speculative sampling: LLMs writing a lot faster using smaller LLMs(blog.dust.tt)
4 points by spolu 3y ago|0 comments
15.
by spolu 4y ago
Hi, this is a fair concern. We're super early and working on a proper privacy policy as we speak. But we also provided some color about how we handle your data on our Discord. Copying it here: ``` The privacy section on the landing REA
16.
by spolu 4y ago
Completely agreed with that. We'll revisit if needed but we don't expect it to be too much of an issue at this stage. Our goal is to gather marginally more usage to learn more about productivity use-cases using LLMs. Here's t
17.
by spolu 4y ago
Dust XP1 runs on a Dust app. You just need a "history" input parameter that gets injected at the right place. Does that make sense?
18.
by spolu 4y ago
The real value will accrue as the assistant become more personalised as users use it.
19.
by spolu 4y ago
Completely agreed! I believe I’ll open source the app in the coming days. This is before anything else an exploration at this stage.
20.
by spolu 4y ago
Some additional context here if interested: https://twitter.com/spolu/status/1602692992191676416
21.
Show HN: XP1 – A GPT-based Assistant with access to the browser Tabs(xp1.dust.tt)
48 points by spolu 4y ago|9 comments
22.
WebGPT reproduced from advanced prompting only (using Dust)(twitter.com)
8 points by spolu 4y ago|2 comments
23.
Show HN: Wedding thank you notes auto-drafted with Dust and GPT3(dust.tt)
2 points by spolu 4y ago|1 comments
24.
by spolu 5y ago
Copilot actually knows a bit of Lean and can be helpful when formalizing stuff. But it does not get the critical feedback we get from the formal systems (as it's not designed for formal systems). It would be interesting to eval copilot
25.
by spolu 5y ago
It takes a lot of CPU (lean side) on top of GPU (neural side) indeed. But technically, when properly parallelized, it takes no more than 1-2h to crack the hardest problem we've cracked so far.
26.
by spolu 5y ago
This is definitely something that we'll be looking into more closely in the future. Our old and busted `gptf` tactic was a good start but we can do much better!
27.
by spolu 5y ago
Sorry, late reply. Indeed settle.network expired and is now owned by another party. Sorry for the ensued confusion.
28.
Decentralized trust graph for online value exchange without a blockchain (2017)(web.archive.org)
123 points by spolu 5y ago|39 comments
29.
by spolu 6y ago
Completely agreed.
30.
by spolu 6y ago
(30k proofs to be precise, valid and test are each 1k)

More ›