triMichael

I highly disagree, though I admit I am one of those towards the top of literacy and cognitive skills. I absolutely enjoy when someone uses LLMs to learn. However, I don't think that's what we'll be seeing.

Look at programming. What's happening there is we are having agents produce 10-100x the code to where people give up on reading it and just hit OK. Look at math. Plenty of people there want to learn from the LLMs and all these crazy new math proofs. But those are the people who will probably be out of a job soon because the money in math is in producing proofs, and that's now an LLM-exclusive job.

I think LLMs are weird in that the first order effects are those of an incredible and amazing tool that anyone can learn things from, but the second order effects are going to be the exact opposite - people giving up on bothering with ideas. Even at my job, which is one of the few that has not gone crazy with AI, anytime I ask anyone a question, all I hear is "have you asked Claude?". There's no communication anymore, no pushback, no struggle. I worry that one day we'll look around and realize we've built an entirely separate second civilization of machines and robots that acts more like a neighboring country than a tool we can use to build humanity up.

-Just my rambling thoughts, not refined by an AI.

I've noticed this as well, I've been comparing it to 4Chan Greentext memes. Everything is a short punchy sentence, and every sentence has the same "weight" to it. There's no flow or rhythm, it's like reading off bullet points.

This is the exact issue I've been having though. I feel like I'm unable to benefit society now. My whole life I've spent learning math and computer science so that I could help other people do things. Now they don't need me, and I feel lost.

Thanks for the correction! I updated my answer to fix that :)

After some back and forth with AI, I think I understand: 1. The Exponential Time Hypothesis being false shows that 3SAT specifically is sub-exponential time 2. Any NP complete problem can be changed into 3SAT in polynomial time

Put these two together and you would think that they would mean that any NP complete problem can be solved in sub-exponential time. But this isn't necessarily true because changing the NP complete problem into 3SAT can increase the number of variables, which can theoretically negate the sub-exponential time.

No, they are completely different. Determinism is the idea that if you have a perfect snapshot of the universe at one time, you could run that snapshot forwards or backwards using the laws of physics and have a complete history of the entire universe. It's a valid scientist idea, but Bell's Inequality experiments pretty much prove it false by showing that either realism or locality is false, and you kind of need both to have determinism.

Meanwhile, superdeterminism is the idea that you can still have realism and locality by claiming that Bell's Inequality experiments are invalid because the universe could somehow "know" you are performing a Bell's Inequality experiment and rig it to get the "wrong" answer. Superdeterminism is equivalent to saying that we can't prove smoking causes lung cancer, because the universe might rig any science experiments to make it look like smoking causes cancer. Anytime you ask why the universe would do that or how you can prove anything in such a universe, they shrug their shoulders. It's a completely BS idea that got traction because specific people pushed it and others don't understand quantum physics enough to know how unserious of an idea it is.

I've had a theory for a while now that music is related to Kolmogorov complexity, or in other words how compressible it is. Random noise that we find unpleasant is very incompressible. On the other hand, simple sound patterns that we find boring such as Do Re Mi is very compressible. What we call music sits in the middle, and is maximizing the amount of signal. In this way, we can view humans as information seeking machines, which is why we like music.

As someone relatively anti-LLM, I fully agree we should run them on our own code for a bug scan. I think the only two scenarios that make sense to avoid it is if you are ethically opposed to them AND releasing something just for fun, or if you are worried about IP issues from them training on your data AND haven't released the product yet. I have something in that second category, my plan is to run an LLM on it immediately before releasing.

I've studied this a lot, and it really isn't a band-aid at all.

Everything aligns perfectly with matter that only interacts with gravity. This means it is invisible and nearly intangible. "That sounds crazy, nothing like that should ever exist" you might say. Except it already does - neutrinos. We just don't think neutrinos quite fit the bill.

Imagine if scientists were trying to identify some sort of sea monster, where cameras have gotten minor glimpses here and there. They aren't sure whether it is a giant squid, or octopus, or whale, or whatever and there's a lot of debate around it. And some non-scientist comes along and says "that's obviously a band-aid, we all know that animals can't exist in the ocean, so maybe you are wrong about what the ocean even is."

The problem here is two-fold. First, the non-scientist is misunderstanding the debate. The scientists actually have a pretty good idea of what's going on, they just lack the evidence to nail down the exact thing they are seeing. Meanwhile the non-scientist thinks that they are completely stumped and need an Einstein level change in direction in their thinking. Second, the non-scientist thinks that nothing acts like what the scientists are looking for. However, that's not true, the "weird" attributes aren't all that weird, they just aren't familiar to the non-scientist.

The irony is that while everyone thinks that dark matter is a band-aid solution and there's not much room for it to be, there is currently an idea that is likely to be a bandaid solution - dark energy. The main idea behind the acceleration of the expansion of the universe is that empty space contains energy, which we call dark energy. This makes logical sense and quantum physics agrees it does. However, any kind of actual measurement of the amount of energy per unit of space that we try just doesn't make sense. The theory doesn't align with our measurements to a vast degree (to where it's known as the worst prediction in the history of physics), and our measurements don't even align with each other. There is room here for a truly novel solution, such as varying amounts of energy per unit of space, or the cosmological principle being incorrect.

As someone who has really struggled with AI recently, I actually think I have an answer to this part: understand the difference between art and science.

Now, I don't mean art as in pictures. Rather, I mean science is objectively correct or incorrect, while art is not. So things like does your code work is a science, and things like your design is an art. In regards to science, AI will do the same things as you, because there's only one "truth". However, in regards to art, it would never do the same things that you do.

One of my favorite games is Tunic. If the developer had tried to one-shot Tunic with an AI, it wouldn't have been Tunic. Maybe it would have been a good game, maybe it would have had a lot of similar ideas. But they wouldn't have been the exact same ideas. Tunic itself would have been forever lost.

So what am I trying to say? Find parts of programming where the path not taken matters. Then it doesn't matter if Claude could one-shot it, because it's one shot will be different than what you do.

Unfortunately, this doesn't fix all of programming. The more science-y parts of programming such as reverse engineering do get affected by AI. But to give you an idea, I'm building a programming language. I toyed with the idea of using LLMs. When I decided not to, I had a problem - my language had become too complex to do it without AI. So I sat down and scrapped the entire thing. Now I have a new programming language that's even better. If I had one shot it with Claude, I wouldn't have had this language. So in the end, doing it by hand mattered.

I'm finding that one of the most difficult parts for me of the AI takeover is the loss of stories.

Here's a good one for you: "Explain the double slit experiment which way variation"

If they say anything about leaving two straight lines, then it fails. Just tried Gemini, and it failed.

This is an extremely common misconception that has spread all throughout the internet, and so it is baked into the training data. The real answer is that there are multiple ways to do which way double slit experiments, but Einstein's thought experiment proves it's impossible for any of them have an interference pattern, as that would violate Heisenburg's Uncertainty Principle.

Somehow, not leaving an interference pattern became twisted into leaving a specific pattern of two lines, which then falsely implies that quantum objects lose their quantum behavior in certain circumstances. The field of quantum physics becomes so much simpler to understand once you realize that all of this is hogwash.

The best reference I can find for where this myth started is a documentary about quantum physics that tries to connect it with mysticism. On the other hand, Wikipedia actually has it correct. In its "which way" section in the double slit experiment page, it correctly says "A well-known thought experiment predicts that if particle detectors are positioned at the slits, showing through which slit a photon goes, the interference pattern will disappear".

I'll add to that: you are more likely to have a good experience if it has a lot of relevant data that it was trained on. You are also more likely to have a good experience if errors don't cause major issues.

So one-shotting a game of Snake should be great (tons of training data, errors are easily caught because it's a small program). Similar with building a lot of web UI front end, or one-shotting a personal project. On the other hand, I haven't been convinced that it's good enough to maintain large codebases or assist with niche topics that are not very well documented.

I'm one of those developers, and I think it makes sense to write documentation for Claude and I have no issues with that.

The point of self-documenting code isn't to get rid of the documentation. Instead, it's to integrate it into the code. This fixes the two biggest problems with documentation. First, that the code will often be updated and documentation left behind, making it useless. And second, that English and code are intertwined in the same document, making you constantly switch mental contexts of how you are reading. So no, I don't consider developers who write documentation to be inferior. If anything, I value documentation even more, and the purpose of self-documenting code is to make the documentation better, not get rid of it.

With Claude, the documentation you make for Claude is just fine as it doesn't run into any of the above two problems. The documentation isn't being left behind, because you add to it instead of the code. And it's not intertwined with code, because you are just writing English for Claude, you are not writing code in-between.

While I haven't had this issue with Gmail, I recently got a new computer and the first two weeks for full of moments like this. It's shocking to me how much we've let popups go rampant on everything. Perhaps the worst offender is Windows update, as it won't even let you use your own computer without clicking through 10 screens refusing all sorts of products they are trying to push on you.