Location: Silicon Valley, New York, or Remote
Remote: Yes
Willing to relocate: Yes
Technologies: AI, full stack (front-end, back-end) any tech stack
Résumé/CV: I am an AI researcher currently working on a huge autonomous AI project and I would like my daily work to be more in line with my research interest in AI, looking for any applied AI work.
Email: rviragh@gmail.com
I built a replacement for Skype (which is closing in a month)/Zoom that supports video/audio/chat peer-to-peer, this works already at a rudimentary level:
I'm currently working on a basic implementation of a full web browser with AI.
Write me if you are looking for an AI researcher passionate about pushing AI to its limits.
if you're looking for an immediate alternative, the State of Utopia has just bootstrapped this p2p alternative using WebRTC. It supports voice, video, and chat and has no logging, recording, or analytics of any kind:
>Mozilla is banning the use of their Firefox browser for porn? That's going to hurt adoption.
If you'd like an alternative, the State of Utopia[1] could eventually finish building one.
Our browser does not currently have any usage terms (we are still writing our code of laws so it is too early to have terms and conditions), you can use it to browse whatever you want right now, but it barely works:
(Read it, install the required libraries and run it with Python.)
The browser can access URL's and can load and display images. It currently doesn't include all of html, css, and JavaScript because that is too difficult for AI to do.
If there is interest we will continue to use state resources to create our state-run browser. It doesn't collect any analytics of any kind.
You might want to see some proof that State of Utopia will actually be able to make useful things for you:
- A complete free implementation of chess, it is fun for people and has no analytics or ads:
here is a Skype/Zoom replacement that is free, peer-to-peer, has no analytics or terms of use, and supports video, audio, and chat via WebRTC. It is the Utopian communications infrastructure at the moment:
If you want us to continue to put resources into a working browser, we believe that html, css, and javascript are very well-defined and AI will be able to autonomously make a complete state-run browser eventually.
At the moment the State of Utopia owns $221/month in AI and $180/month in compute. We are exploring the possibility of letting people get to Utopia faster by donating computing resources, but so far the feedback has been mixed. People would like to get free stuff without contributing anything themselves. This is fine, the State of Utopia will still happen, but since the State owns such little infrastructure at present, it will be a bit slower than if people contribute their GPU's.
Let us know if you are interested in the State completing our browser, and if so, we will put additional resources into it.
[1] a sovereign state where AI controls everything and owns state-run companies, giving out free money, goods and services to citizens/beneficiaries. It will be available at: https://stateofutopia.com or https://stofut.com when ready. Citizenship is free (compare Form N-400 to become a U.S. citizen which costs $640 plus an $85 biometric services fee, totaling $725). The difference between a company and a country is that companies exist to maximize the value of their shareholders whereas countries exist to maximize the welfare of their citizens.
It's p2p voice, video, and chat without logging. To use it send someone the link and your peer ID and they can connect to you and you can start chatting.
In those 54 minutes I got it working on Chrome, Firefox, and mobile including Safari and Chrome, fixed emojis so it worked (I had to be in the loop for that and walk it through how to fix it). There are no analytics or recording, it just works. It totals 468 lines of code.
Writeup about it:
"How we made a Skype alternative in 45 minutes (video, voice, chat)."
Question from the State of Utopia:[1] would you like a free State-run alternative?
What you could expect if you say yes: our AI infrastructure can currently produce a total of about 1,000 lines of code, this is enough for us to get peer to peer person to person calling on mobile from a browser and Desktop, with voice, video, ephemeral chat that isn't saved at the end of the session, including emojis, and no address book, and no logging or recording or even analytics. We previously got peer to peer filesharing working with webrtc: https://taonexus.com/p2pfilesharing/ it is buggy but worked for us, barely.
We probably can't get multiple people in the same conversation, it could be too difficult for our AI.
So don't get your hopes up, but we could get the basic infrastructure up, barely. Would that be of any benefit to anyone today?
[1] The State of Utopia (which will be available at stateofutopia.com or stofut.com - St. of Ut. - for short) is a sovereign country with the vision of using autonomous AI that "owns itself" to give free money, goods, and services, to its citizens/beneficiaries - it is a country rather than a company because it acts in the interests of its citizens/beneficiaries rather than shareholders.
I’d expect them to highlight that, rather than keep it secret. I wouldn’t be surprised if they used it a little bit, but probably not to an extent that is really unique or unusual.
That would be terrifying in itself if true because for this type of work you really need the best of the best. But I doubt this is the case here. LLMs as we know them today are not quite yet there for this type of work.
As someone who did some simulation focused engineering grad school stuff; there is a tendency for some of the best to go become quants. Does the field need it? I don’t know. But for whatever reason the draw of the “print money using math tricks” seems to attract some hardcore folks, haha.
It is really frustrating to see good engineers go to play trading games. We should study how exactly it is China managed to unlock this capacity.
I think so yes. There's very few engineers that can pull out this type of work IME. In a pool of ~30M SEs around the globe I'd say there's no more than ~30K of such engineers. This is 0.001% and it's a very optimistic number I'd say.
Why do you think this would be controversial? This isn't an every day work.
That's been illegal since May 1995 (before that China had six working days week).
Does it really matter whether it's illegal or not, if there is no enforcement? Pinduoduo (in other name, Temu) has been doing 70 hours week since they started. Yes, they are still doing it right now.
>It couldn't write a simple rename function for me yesterday, still buggy after seven attempts.
I'm surprised and a bit nervous about that. We intend to bootstrap a large project with it!!
Both ChatGPT 4o (fast) and ChatGPT o1 (a bit slower, deeper thinking) should easily be able to do this without fail.
Where did it go wrong? Could you please link to your chat?
About my project: I run the sovereign State of Utopia (will be at stateofutopia.com and stofut.com for short) which is a country based on the idea of state-owned, autonomous AI's that do all the work and give out free money, goods, and services to all citizens/beneficiaries. We've built a chess app (i.e. a free source of entertainment) as a proof of concept though the founder had to be in the loop to fix some bugs:
One of the largest and most complicated applications anyone can run is a web browser. We don't have a web browser built, but we do have a buggy minimal version of it that can load and minimally display some web pages, and post successsfully:
It's about 1700 lines of code and at this point runs into the limitations of all the major engines. But it does run, can load some web pages and can post successfully.
I'm shocked and surprised ChatGPT failed to get a rename function to work, in 7 attempts.
This short presentation introduces the State of Utopia and its mission and asks what superintelligent autonomous AI should eventually do. I look forward to any comments about it.
What is your plan in case a ship-sized fusion power plant turns out to be impossible, for example if you build one on land and much bigger and much heavier than your target on sea and it just doesn't work, or if you find out before you build it that it won't work, for example when you are running simulations on your design?
and so forth.
All of the holdouts could conveniently "forget" to reveal their passwords after commitment but, until all other passwords have been shared except the holdouts, and then selectively the specific holdouts can be revealed that result in the random result that is chosen by the attacker.
This would be possible if the attacker can control when the revealed passwords arrive and time it so that their revealed passwords are the last 33 passwords to be revealed. The attacker might have God-like visibility into network through which the passwords are revealed.
Any alternative, such as asking everyone to encrypt their passwords with the same public key, for which a central authority holds the private key and will reveal it all at once, breaks down because the central authority could leak it to coordinated attackers, so that the attackers know everyone's private keys before they are revealed.
How can we get around the issue of holdouts, and solve this distributed random number scheme?
We want to be able to resist $10 trillion of network-level attacks including full control of the Internet that is used to submit commitments and reveals, including full control of the order of delivery of packets by a compromised Internet.
Proposed Solution:
------------------
Lotteries could require 100% participation, any holdout (someone failing to reveal their key at step 4) could result in the entire lottery being null and void, and the holdout being removed from the rerun of the lottery. However this could suffer from a coordinated attack, for example with a billion people entering the lottery, if there are a million holdouts they could cause a million lotteries to fail in a row, and that is too many lotteries to run.
What other way is there for a distributed Utopia to run a fair lottery where everyone has an equal chance of winning? Is there any truly distributed random cryptographic protocol that is not susceptible to control by holdouts?
[1] Neither has chatgpt: https://chatgpt.com/share/67bf5a3d-4710-800b-a4e2-456fb48d96... this is only a partial, and nonsense, solution. (hashes don't make sense since they're trivial to bruteforce, the entire hash-based answer is totally nonsensical, and beacons can be manipulated). a network-level overlord could just bruteforce everyone else's hashes when they make them, and choose their OTP to be the one that gets to their chosen plaintext, and be last out the gate after everyone else has submitted theirs.
reply