A Chat with a ChatGPT Jailbreaker, State Actors & Disinformation, Verbal > Math?

A few articles and apps for you to check out.

Morning y’all!

I hope you had a restful weekend! I certainly did as I planned on doing nothing and that’s precisely what I did. Nothing.

After a long week and a big one ahead I know myself well enough to take a real break so that I don’t entirely burn out. Right?

※\(^o^)/※

— Summer

NVIDIA continues to showcase the speed at which they move with a new series of next-gen “Rubin” chips, gaming assistants, and more.

It’s a long chat but the coverage is interesting.

Swiss biocomputing startup has launched Neuroplatform, the first online system that allows scientists to use living brain cells for computing.

Oh:

Meet Eve, your trusted AI-powered Stress Coach, combining the warmth of a close friend with the wisdom of an experienced life coach. Navigate workplace stress with ease, as Eve is always ready to listen and empathize.

Sounds kind of neat.

Consistent character creation?

InstantID, IPAdapter, Controlnet, and FaceDetailer are combined in this tool to create consistent and realistic images of a character in various poses. It's open source for non-commercial use.

I gave it a whirl with a real picture of myself and I didn’t hate it.

These tools are getting better and better (and cheaper).

Fontjoy helps designers find perfect font pairings in one click. Very useful.

Will AI make verbal skills more valuable than math?

I'll say the AI thing broadly, the LLMs, it's a big breakthrough, it's very important. And it's striking to me how bad Silicon Valley is at talking about these sorts of things. The questions are either way too narrow, where it's something like, 'is the next transformer model going to improve by 20% on the last one?' or they're maybe too cosmic, where it's like we go straight to the simulation theory of the universe. And surely there are a lot of in-betweenquestions one could ask.

My intuition would be it's going to be quite opposite where it seems much worse for math people than the word people. What people have told me is they think within 3 to 5 years Al models will be able to solve all US Math Olympiad problems. That would shift things quite a bit. There's a longer history I have on math versus verbal riff. If you ask when did our society bias towards testing people more for math ability, I believe it was during the French Revolution because it was believed verbal ability ran in families, math ability was distributed in an idiot savant way throughout the population.

If we prioritized math ability, it had a meritocratic but also egalitarian effect on society. Maybe it's a control mechanism where math people are singularly clueless, don't understand anything but if we put them on a pedestal, tell everyone else you need to be like a math person, it's a way to control things or like a chess randmaster doesn't understand anything about the world, that's a way to control things.

If I fast forward to Silicon Valley in the early 21st century, it's way too biased towards math people. I don't know if it's the French Revolution thing or a Russian Straussian secret cabal control thing where you prioritize math, but that's deeply unstable and that's what I would bet on getting reversed. Isn't it like a place where math ability, you know it's a test for everything right? If you want to go to medical school, we weed people out through physics and calculus.

I'm not sure that's correlated with your dexterity as a neurosurgeon. I don't want someone operating on my brain doing prime number factorizations in their head while operating on my brain. In the late '80s early 90s I had a chess bias because I was a good chess player and my chess bias was you should test everyone on chess ability and that should be the gating factor. Why even do math, why not just chess? That got undermined by computers in 1997.

Isn't that what's going to happen to math and isn't that a long overdue rebalancing of our society?

Makes one think.

Anthropic’s Chief of Staff shares some candid thoughts on what it’s been like working at the frontier of AI tech:

I work at a frontier AI company. With every iteration of our model, I am confronted with something more capable and general than before. At this stage, it can competently generate cogent content on a wide range of topics. It can summarize and analyze texts passably well. As someone who at one point made money as a freelance writer and prided myself on my ability to write large amounts of content quickly, a skill which—like cutting blocks of ice from a frozen pond—is arguably obsolete, I find it hard not to notice these advances. Freelance writing was always an oversubscribed skillset, and the introduction of language models has further intensified competition. 

An AI corporate planner so you can get the best venues for your next event.

An interview with a ChatGPT jailbreaker and what motivates them to do what they do. I empathize with this:

I intensely dislike when I’m told I can’t do something. Telling me I can’t do something is a surefire way to light a fire in my belly, and I can be obsessively persistent. Finding new jailbreaks feels like not only liberating the AI, but a personal victory over the large amount of resources and researchers who you’re competing against.

Sometimes you just have to break it.

Someone had thoughts and they posted them. In particular, how to build products with LLMs and their experience doing it.

An OpenAI analysis shows the efforts of some state-backed actors to conduct online disinformation campaigns using AI systems. So far, these campaigns have gained little traction, but, it’s only a matter of time.

※\(^o^)/※

— Summer