Hacker News Reader

monkpiton Doing a Machine Learning PhD While Working in Japan

Agentless passive, hedging, zombie nouns. It’s the type of writing that’s meant to meet a word count.

I think a scooter or a bus is a more apt comparison than a steam engine. The scooter and bus can both get the job done with some acceptable trade-offs, depending on your circumstances and what you’re willing to accept.

However, there are many use cases where they aren’t the right tool for the job.

monkpiton Coding is not solved

Yes, I fail to see anything meaningful. If you move the goalposts and say “I have invented a human that cannot be convinced in any way to give a wrong answer” then what’s the point of that in this discussion, really?

And you seem confused about how an LLM works and what it is - “the intelligence of all humanity” - not how it works. You’re debating using 2 imaginary things you created.

monkpiton Coding is not solved

What’s your point though, really? “You can train a model to say things that are objectively wrong”? You can do the same with a human.

monkpiton Coding is not solved

> You see this already, LLMs are a lot more reliable in statically typed languages with strong memory guarantees (like typescript or rust) than in weaker languages.

It’s common advice to wire in deterministic feedback to your workflow with LLMs - static languages aren’t inherently better for LLMs, it’s that LLMs produce better code when given deterministic feedback, such as compiler results.

monkpiton Creatine uptake enhances antitumor immunity

Some people report less GI distress with creatine HCL but it’s much more expensive.

The GI issues are caused by the creatine absorbing water. This causes your gut to have more water than usual as it follows the creatine through the works… which causes… urgency.

The effect diminishes after you’ve been taking a regular dose for a while, even with just monogydrate (not HCL).

Some people say HCL causes less distress because it doesn’t draw water as much. Others argue that defeats the purpose of taking it.

Monohydrate is studied to death and proven effective, HCL not so much.

monkpiton A brief history of Windows scroll bar shortcuts

Yeah, I’m missing something here… didn’t he say that it goes down by a page if you click there? I always thought it just took you to that spot.

monkpiton What California is learning from solar panels built over irrigation canals

Apparently the definition of clear is debatable here? It’s ridiculous to think a mechanic would pay people to damage gas tanks in the hopes of drumming up business.

monkpiton What California is learning from solar panels built over irrigation canals

Never attribute to malice… etc.

monkpiton You can defeat the Dream Devourer from Chrono Trigger using an int overflow

The point is that it doesn’t matter, they enjoyed it. The fun police weren’t invited to do an investigation, no matter what scenario or analogy is applied.

Saying “you didn’t like it right” is gatekeeping.

monkpiton You can defeat the Dream Devourer from Chrono Trigger using an int overflow

The analogy to a book doesn’t fit anyways.

The closest thing would be telling someone they’re reading a choose-your-own-adventure book wrong somehow, and even then it’s a stretch.

monkpiton You can defeat the Dream Devourer from Chrono Trigger using an int overflow

Why announce that you’re incapable of allowing someone to hold an opinion that doesn’t match yours?

Would it be so wrong for someone to enjoy something in a different way than you do? Or should they be chastised for enjoying it the wrong way?

You’re welcome to ask what brings them enjoyment, to broaden your views, instead of dismissing them.

monkpiton How OpenAI Used Its Own LLMs to Design Its Jalapeño Chip

Or cyber…

monkpiton Introducing System One Models and Jev

404?

monkpiton Introducing System One Models and Jev

To be fair - you’re crafting a deliberately bad model input for a contrived example.

monkpiton Chopping up books when they're physically too big

Why can’t someone else have a different opinion than you, even if it doesn’t make sense to you?

monkpiton Cartesian – AI 3D Modeling for Design

Are you hooking it up to blender via some MCP or are you just creating files and opening in blender?

monkpiton OpenAI bots knew about the RubyGems caching vulnerability

Intent might be what’s being discussed but intent is, for the most part, legally irrelevant. It might make the difference in the degree of a murder charge, or maybe manslaughter, or criminal negligence, but it doesn’t get you off the hook.

monkpiton OpenAI agents carried out an undisclosed attack on RubyGems

IMO the argument about anthropomorphizing misses the point - what most comments that talk about anthropomorphizing really want to talk about is accountability. It’s impossible to hold an LLM accountable, and in rare cases where people do (that guy who got his prod db deleted) it comes off out of touch. The rest, though, is basically inconsequential - whether you attribute emotions or agency to the LLM doesn’t really affect much if you accept that it can’t be held accountable (but the human can).

monkpiton Eating Fruit Skins

There are old mushroom hunters

and there are bold mushroom hunters

but there are no old bold mushroom hunters.

monkpiton Tao: Open math problems being non-renewably mined by AI

This is my question too. If we are all just going “well that sucks” after AI solves this problem, why did anyone care about the problem being solved in the first place?

Is the bummer that we got a solution we didn’t want - that navier-stokes is not always applicable or something, but we hoped it was?

monkpiton Tao: Open math problems being non-renewably mined by AI

To what end?

monkpiton I-have-ADHD: A skill to stop coding agents from burying the answer

Have people done that? Or are they just meat puppets?

monkpiton Holden's Lightning Flight

> Holden ran down the runway, narrowly missing a crossing fuel bowser

TIL Bowser is an actual word, huh.

monkpiton Muse Spark 1.3

Would you care to discuss the topic, or just throw grenades? Surely you can come up with something more substantive than this

monkpiton Muse Spark 1.3

You’d expect that, wouldn’t you? But, alas…

monkpiton Muse Spark 1.3

Would be cool if there was a benchmark to evaluate the “tool-like” quality of a model - its capability to quickly, cheaply, accurately, do exactly as it is asked.

monkpiton Claude Fable 5.1 and Claude Mythos 5.1

> Considering they put Opus 5 (High) at the same level as Fable 5 (Max), I don't trust it that much.

Have you used both? I’ve never experienced any seemingly greater level of intelligence from Fable 5 over Opus.