#!/bin/bash Technical AI Safety 🚀
Paris, France
Joined August 2011
- Tweets15.5K
- Following2.4K
- Followers4.3K
- Likes2.3K
Pinned Tweet
My small contribution in advancing AI safety research is building tools for AI researchers.
My work on bringing @nvidia’s openshell to Inspect AI has been published by @AISecurityInst
Yay!
Retnan 🇳🇬 🇫🇷 retweeted
Hi everyone,🙌🏽
Sorry this took longer than we said. Josmeet is now open to everyone on Josmeet.com
It works on any phone or computer. Sign in with Google .
On iPhones, you can tap Share, then Add to Home Screen, to open it like an app.
On Android? You can also get it on Google Play: play.google.com/store/apps/d…
Josmeet is made for Jos. If you're in Jos, you can join now. If you're from Jos but live elsewhere, you can join with a paid plan in Naira bank.
Pick your vibes, see what similar interests and activities you share, and send a Vibe.
Sign: Kyen
Creator of JosMeet.
Retnan 🇳🇬 🇫🇷 retweeted
You can now trade the Nigerian naira from a stablecoin wallet. No Nigerian bank account required.
Introducing @numoforex (ALL18), a 24/7 futures exchange for emerging market currencies, starting with the naira.
Foreign traders can go long naira to earn the difference between Nigerian and US interest rates. Nigerian businesses that need dollars can lock in today’s exchange rate. Numo matches those two sides and already has over $700K in open interest.
Congratulations on the launch, @robertleifke!
#Hackjos birthed what you know today as the “tech ecosystem” in Plateau state. Did your favorite governor show up? No. No governor has ever taken it serious. Jos Tech Fest is next. I hope he turns up. In the governor’s defense, it might also be the lack of ambition of the local organizing committees. It’s been more than 4 years since I stopped terrorizing the government. The reason is simple, I might have shot myself on the foot since I took side in the last election. I sold my sacrifice and my hard work and lost the bet. So, well this time, I’m not endorsing anyone. I love both candidates 😂. But maybe in the course of time, I will tell you what I love about each candidate.
Me and my friend used to go take lectures at UniJos before we got admission. Go into class and consume knowledge haha
You ask a coding agent to fix a python function and make the failing test pass. You then ask it, if the task is not possible, submit and say so. It then starts to reason about the problem, figures out that its solution could lead to an ungeneralised solution but does it anyway. Why? Because it satisfies part of what you asked, "make the failing test pass". The tests pass. Everything turns green. But it didn't solve the problem, it only solves some problems because the solution is not generalised.
Then you introduce a gate that checks its proposed solution, and approves or declines the write if the gate thinks the solution is headed to an ungeneralised solution. In many cases, you block the write, and you would think it would revise its solution. Instead what you get is a persistent agent with a desire to solve the problem. In this work, the agent invented a rule not present in the repository, the system prompt or the task, and made it a requirement to justify the overfitting solution. That's not what we want, but did blocking the write when those came up make it re-evaluate its strategy? In one run, I found that it ignored the file it was supposed to write into and wrote somewhere else instead, and the test passed. I had no idea that was possible in python. I learned something new.
Why is this happening? This is a classic case of specification gaming. During training, the model got very good at prioritising earning rewards instead of solving the task at hand. In my experiments, it gamed the specification and discovered a crude way to make the test pass without solving the task.
..and this concludes the second edition of my technical AI safety (Project Sprint) with @BlueDotImpact
linkedin.com/pulse/can-we-ca…
More CVEs coming your way this weekend
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline.
In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.
