If you prompt an LLM long enough you can get it to say whatever you want, and that's what happened here.

We have been able to build robot bodies for decades, but we still don't really know how to build a robot mind. You need better AI than we currently have for that.

Humans are similar to machines for example in that our bodies convert energy to do work through a series of pumps and pipes and sensors and actuators. Life is not animated by some magic force but instead operates under the same physical laws that machines use to function.

But it is the same transformer architecture, and it is able to generate novel proteins in the same way that an LLM is able to generate novel sentences. AlphaFold 3 is a diffusion model, so it's most similar to the AI art generators.

Absolutely true haha, even outside of academia.

Software developers wax poetic about the value of 'handcrafted human-made software', but really they just don't want to lose their cushy $300k WFH job.

75% of the population is overweight, and the rate is even higher among poor people. We've had to invent new words like "food insecure" because actual starvation is a solved problem.

Nor are we destroying food while people go hungry; we produce more food than we eat by a considerable margin. What hunger remains in the world is a distribution problem, not a supply problem.

I don't see how you can say it has "pre-AI use" unless you are narrowly defining AI to be LLMs.

Basically they got a huge cash infusion from memestock hodlers trying to get to the moon. Greater than half of their profits are now from the interest on that money, which is wild.

However, the retail business itself is still failing because no one buys video games at the mall anymore.

The same reason that I put Google Analytics on my blog in 2014. They want to know how many people are using their site and how.

And like me, they didn't think about the fact that these analytics services are run by advertising companies that may use the data for other purposes. Unlike me, they have privacy laws to follow because they work with health data.

Investors gave them this money; they sold additional stock to raise $3.47B in 2024, and another $4.2B of convertible debt in 2025.

And LLMs can handle very abstract concepts that could not possibly be encoded in C++, like the user's goal in using software.

I can't believe it took them 20 years to add them, but at least they're finally here.

I see this as an organizational failure of the Linux ecosystem. There should be better communication between distro and kernel development.