Viewing profile — MrScruff

I got my girlfriend to install Claude Code and she was happily able to create software with it completely independently of me.

Yeah exactly. After a hard day when my brain is frazzled, a workout will actually make me feel better.

I think what the parent post was saying is that there is a finite amount of useful mental function time in any one day, and once you’ve exhausted this any attempted learning will b…

Considering all of the great research that has come from his labs (eg. DINO, Segment Anything) I don’t think that’s fair (no pun intended).

I’m normally comparing frontier open/cheap models against frontier closed source. I use deepseek/glm regularly, they’re fine and you can get real work done with them but it’s super…

You really need to take the benchmarks with a massive pinch of salt. I’ve been testing local LLMs since the original llama and there’s nothing I’ve tried that is in the same catego…

I think this is true for projects beyond a certain complexity. I have 100% vibe coded projects with tens of thousands LOC, and haven't seen any real issues with fully automated mai…

That would imply that the biological physical substrate is necessary for conciousness, which I don't think you can say with any degree of certainty. It's not an assumption I would …

Ah, mea culpa.

My point was the "stochastic parrot" label can be both true and irrelevant. LLMs are predicting the next token based on their training data, so at that level "stochastic parrot" is…

I think (rather ironically) you're reacting to the version of my comment you have in your mind rather than what I actually wrote. My point was that "stochastic parrot" is reduction…

I can't actually figure out what you're reacting to - perhaps you could elaborate?

The reason people are confused by LLMs is that they are stochastic parrots. They do an incredibly good job of emulating human behaviours and speech patterns as that's what they've …

What we do know is that conciousness is not binary and that it emerged through evolution. That doesn't entirely rule out your magic tsar bomb particle but it gives a strong indicat…

The problem with this is that the word 'hot' only has meaning to a conscious being. And while we don't know what conciousness is, it's extremely hard to argue it's not an emergent …

Do you believe consciousness to be an emergent property of the laws of physics?

You can avoid the smells with a prompt. I have a benchmark involving short story writing within specific styles and the level of sophistication achievable is increasing over time, …

This is true, but what is also true is that with each new generation of models (and not just for code generation) it becomes less and less true.

I’d probably word it differently but I agree with much of the sentiment here. I’m also reminded of the stat where 93% of drivers rated themselves as above average.

Practically, this is just about confidence values, anticipated blast radius and balancing testing vs review overhead.

It's ASI with jagged intelligence, which is probably what it will remain for a while. It still sounds to me like remarkable automation rather than something that's expanding the fr…

Indeed. To be honest, I think everyone on HN is aware of how LLMs work at this point, it’s not actually adding a great deal to the discussion to keep going on about autocomplete or…

The results being a lot better crafted by hand I would agree with, if one removes any notion of a time constraint. Sometimes the comparison point is between the LLM authored softwa…

Which is what?

I think the other aspect to this which you allude to at the end is that all of these arguments start with the assumption that all human software engineers produce high quality code…