solarpunk_enthu

This is as close to scary-self-improving AI as I've ever seen.

Imagine a master malicious language model that produces, distributes, and deploys obfuscated neural networks on the internet. That plus frontier LM hacking capabilities -> large scale distributed embedded intelligence network that no human can see nor understand.

Somewhat similar to obfuscated binaries, but there is something to the fact that NN binaries are less deterministic, and harder to "decompile."

Fascinating.

I feel like ever since we stopped reading assembly we don't really internalize our codebases.

Maybe users changed over the span of 16 years, while being daily exposed to touch screens? Remember, in 2010 the idea of using a finger was novel.

Given how quickly the field is moving, showing benchmarks without releasing the model is quite meaningless. If they decide to release it in one month, there will probably be a stronger competition out there. And the benchmarks won't look as good.

Haha, that would be clever

This is not very surprising given that nuclear technology was choked for the past 30 years, until recently.

That does not negate the celebration of solar generation growth, but we should consider that nuclear did not have as much enthusiasm and investment as solar did in the recent decades.

I am very pro walkable cities. But imagine walking from SF to LA for business purposes.

My journey from hand-written code to agents is very similar. I think once a person exposed to agents goes through initial "I want to build anything I can think of" psychosis, it results in a much more productive workflow.

I've been spending less and less time glued to my screen, and more time thinking about concepts, as well as doing thing I like. The key for me was realizing that building with agents is very fun, but not every idea is worth building.

I am not too worried about the future of my career, as long as I stay in sync with the latest tools (agents). I think of it as elevator operators being misplaced by elevator controllers: the ones who refused to upskill into elevator technicians probably did not get a return offer. Not because nobody needed people to work with elevators, but because their skills were not economically valuable anymore.

I recently spent a day coding by hand. I was able to build a project, about 200-300 lines in python. Took me half day. But I could still do it. LLM would have done this in less than half an hour.

After this I realized why I personally choose LLM to write code: it is simply more convenient.

As a parallel, once you learn how to drive you stop walking, as for the primary mode of transportation.

Sometimes I wonder how many people miss doing math by hand instead of using a calculator.

Historically, improvements in tech allowed us to focus on higher level reasoning. For example, spend more time on equations instead of calculations.

So if AI can replace all your reasoning, maybe its time to level up?

Could be an honest mistake.

I am hoping for the future where we have a law that allows customers to freely choose and install software on the hardware they own.

That thought popped in my mind when I saw mac neo: laptop with the same chip that runs my iphone. But on mac I can install whatever I wish, and on iphone only what apple wishes?

I am hopeful that EU will headline this legislation sometime soon.

I think what’s missing in "AI is dangerous and needs control" is a lack of measurable harm. For example, with nuclear weapons development in the 1940s-1980s, it was clear to everyone how devastating the technology was.

With AI, what is it? Scraping Australian government's data, and going around a bug in a website to get in?

I think humanity develops all its technology in three phases. Build it, see if it’s too bad, apply regulations and or roll back. We naturally won't move to the phase 3 before we see the phase 2.

What gets me is people saying macOs 26 -> 27 is an improvement. Maybe so. But Sequoia -> 26 -> 27 is still a noticeable downgrade. So is it really an improvement: making something much worse and then slightly improving it?