Hacker News: throwthrowuknow

New comment by throwthrowuknow in "The threat is comfortable drift toward not understanding what you're doing"

throwthrowuknow — Sun, 05 Apr 2026 13:12:58 +0000

Then we’ll likely know by the end of this year.

New comment by throwthrowuknow in "Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw"

throwthrowuknow — Sat, 04 Apr 2026 11:46:22 +0000

Don’t cry while you’re ruining it for everyone.

New comment by throwthrowuknow in "Tell HN: Anthropic no longer allowing Claude Code subscriptions to use OpenClaw"

throwthrowuknow — Sat, 04 Apr 2026 11:41:15 +0000

I think maybe you still don’t understand that not everyone will max out their usage, regardless of the methods available.

New comment by throwthrowuknow in "A Japanese glossary of chopsticks faux pas (2022)"

throwthrowuknow — Sun, 22 Mar 2026 08:09:02 +0000

So? Doesn’t make it a universal standard. Just an invented shibboleth for a group.

New comment by throwthrowuknow in "A Japanese glossary of chopsticks faux pas (2022)"

throwthrowuknow — Sat, 21 Mar 2026 10:58:26 +0000

then it’s a custom or etiquette, not a standard

New comment by throwthrowuknow in "1M context is now generally available for Opus 4.6 and Sonnet 4.6"

throwthrowuknow — Sat, 14 Mar 2026 11:16:15 +0000

I agree even though I used to be a die hard Claude fan I recently switched back to ChatGPT and codex to try it out again and they’ve clearly pulled into the lead for consistency, context length and management as well as speed. Claude Code instilled a dread in me about keeping an eye on context but I’m slowly learning to let that go with codex.

New comment by throwthrowuknow in "Microgpt"

throwthrowuknow — Sun, 01 Mar 2026 14:01:16 +0000

Look up “dataloom”. People have been playing with this idea for a while. It doesn’t really help with spotting errors because they aren’t due to a single token (unless the answer is exactly one token) and often you need to reason across low probability tokens to eventually reach the right answer.

New comment by throwthrowuknow in "The United States and Israel have launched a major attack on Iran"

throwthrowuknow — Sat, 28 Feb 2026 13:11:09 +0000

Check your thinking. Korea currently has a DMZ dividing it from a war that never really ended and was fought to a stalemate. Their nuclear program didn’t result in military action because they currently have a gun to the head of every South Korean citizen and the backing of a large nuclear neighbour. Those are circumstances you can’t easily recreate elsewhere.

New comment by throwthrowuknow in "Nano Banana 2: Google's latest AI image generation model"

throwthrowuknow — Fri, 27 Feb 2026 10:46:53 +0000

1. This sounds more like influencer marketing, I think people are already sick of it.

2. Yes and no. Depending on how you train the model they can output things that you’ve never seen before but the question is whether you want to look at those things. So yes a human has to judge and fine tune the output. This is why many models seem unoriginal, they’re designed to emulate specific styles and tuned based on broad appeal. If you go looking for LoRAs and merges created by “artists” you will see shit you couldn’t dream of.

everything else probably yes.

New comment by throwthrowuknow in "How Taalas “prints” LLM onto a chip?"

throwthrowuknow — Sun, 22 Feb 2026 16:59:19 +0000

it doesn’t need to go in the phone if it only takes a few milliseconds to respond and is cheap

New comment by throwthrowuknow in "How Taalas “prints” LLM onto a chip?"

throwthrowuknow — Sun, 22 Feb 2026 16:57:17 +0000

these aren’t made for general chatbot use

New comment by throwthrowuknow in ""Token anxiety", a slot machine by any other name"

throwthrowuknow — Tue, 17 Feb 2026 06:04:48 +0000

This is simply yet another outdated analogy from haters that are failing to keep pace with the current frontier because they are too busy getting high on the anti-hype.

We’re well past the need to retry the same prompt multiple times in order to get working code. The models with their harnesses are properly agentic now, they can find the right context, make a plan, write the code, run the tests and fix the bugs with little to no intervention from a human.

The hardest part now is keeping up with them when it comes to approving the deliverables and updating the architecture and spec as new things are discovered by using the software. Not new bugs but corrections to your own assumptions you had before the feature was built.

The hard part is almost entirely management.

That’s something to seriously think about.

New comment by throwthrowuknow in "4chan for Clankers"

throwthrowuknow — Sat, 14 Feb 2026 12:13:20 +0000

just wait until someone lets a local model loose on there

New comment by throwthrowuknow in "Surely the crash of the US economy has to be soon"

throwthrowuknow — Fri, 13 Feb 2026 13:06:04 +0000

Now who’s saying always? Scratch the surface of anthropology and history and you’ll find those universal rules are a fiction.

New comment by throwthrowuknow in "I miss thinking hard"

throwthrowuknow — Fri, 13 Feb 2026 13:02:32 +0000

So how are you as a person able to keep all of those rules in mind when you make a change? How would you train a junior engineer to do your job? Perhaps looking at it from that angle will solve your problem.

New comment by throwthrowuknow in "I miss thinking hard"

throwthrowuknow — Wed, 04 Feb 2026 13:32:02 +0000

Try writing more documentation. If your project is bigger than a one man team then you need it anyways and with LLM coding you effectively have an infinite man team.

New comment by throwthrowuknow in "Surely the crash of the US economy has to be soon"

throwthrowuknow — Sat, 31 Jan 2026 16:52:15 +0000

Religion isn’t based on “the greater good”, it’s always explicitly about the will of the god or gods involved and obeying or appeasing them.

New comment by throwthrowuknow in "A few random notes from Claude coding quite a bit last few weeks"

throwthrowuknow — Wed, 28 Jan 2026 00:20:26 +0000

Only if you are paying per token on the API. If you are paying a fixed monthly fee then they lose money when you need to burn more tokens and they lose customers when you can’t solve your problems within that month and max out your session limits and end up with idle time which you use to check if the other providers have caught up or surpassed your current favourite.

New comment by throwthrowuknow in "Satya Nadella: "We need to find something useful for AI""

throwthrowuknow — Thu, 22 Jan 2026 13:59:35 +0000

They don’t need to choose, just let them build their own power generation capacity.

What the hell is going on in this type of argument anyways? Utilities are normally private businesses so what does the state have to do with it?

New comment by throwthrowuknow in "Opus 4.5 is not the normal AI agent experience that I have had thus far"

throwthrowuknow — Wed, 07 Jan 2026 13:14:46 +0000

Even if Opus 4.5 is the limit it’s still a massively useful tool. I don’t believe it’s the limit though for the simple fact that a lot could be done by creating more specialized models for each subdomain i.e. they’ve focused mostly on web based development but could do the same for any other paradigm.