Hacker News: shomp

New comment by shomp in "Microsoft starts canceling Claude Code licenses"

shomp — Sun, 24 May 2026 18:20:14 +0000

So you're making a programming language... and you don't want to read code. Have I got the gist of it?

New comment by shomp in "Ask HN: How did you find PMF?"

shomp — Sun, 24 May 2026 18:14:21 +0000

Sell a solution before you have a product.

New comment by shomp in "Microsoft starts canceling Claude Code licenses"

shomp — Sat, 23 May 2026 15:06:45 +0000

You review the code on github also from your phone?

New comment by shomp in "Microsoft starts canceling Claude Code licenses"

shomp — Sat, 23 May 2026 14:48:45 +0000

My "mental scratchpad" needs to be as sharp as possible to maximize my intelligence. I think of the LLM as a scratchpad for my thinking, I hope the Anthropic team can see this.

New comment by shomp in "Microsoft starts canceling Claude Code licenses"

shomp — Sat, 23 May 2026 14:46:10 +0000

When you say "code on your phone" ... you don't mean what I think you mean do you? Like, are you actually using your phone to make code commits?

New comment by shomp in "How to convert between wealth and income tax"

shomp — Fri, 22 May 2026 23:29:16 +0000

Oh he understands the reasoning. He is one of the people who would be subject to the tax :) The wealthy writing essays on how to not infringe upon their wealth is nothing new, although it did go out of favor in the Carnegie-Rockefeller era when they competed to see who could do more public-good with their immense fortunes before their shoving off the mortal coil. Looks like the ultrawealthy currying sentiment to preserve ultrawealthy wealth at any expense is back in vogue.

New comment by shomp in "'You can hear me now or pay me later' Music exec tells graduates booing AI"

shomp — Tue, 19 May 2026 16:36:38 +0000

The kids are alright

New comment by shomp in "Malta gives citizens a paid version of ChatGPT Plus for free"

shomp — Sat, 16 May 2026 19:33:56 +0000

GPT is unreliable now after government acquisition. This is steps backwards

New comment by shomp in "Thousands todo apps, but none allows one-click collaboration with my grandma"

shomp — Sat, 16 May 2026 19:33:03 +0000

This is really cool! You did a great job with this. I'm glad it's something you're actively using, that's the hallmark of a good product -- users :D

A teeny tiny nit-pick, the bottleneck diagram ... frustrates me. I think the intent is you want to depict the fact that different [classes of] users experience friction with various obstacles ? I don't know, I personally found that it took away from your otherwise very cool article. A live demo would add to the intrigue and fun aspect of your write-up, although I don't know if you are interesting in making a "massively multiplayer online dotolist" ;D

New comment by shomp in "For thirty years I programmed with Phish on, every day"

shomp — Sun, 03 May 2026 19:20:55 +0000

"the higher up the corporate hierarchy, the less music"

New comment by shomp in "For thirty years I programmed with Phish on, every day"

shomp — Sun, 03 May 2026 19:17:03 +0000

If you can't change the work, change the genre

New comment by shomp in "For thirty years I programmed with Phish on, every day"

shomp — Sun, 03 May 2026 19:16:00 +0000

If the LLM behaved more like a programmer-in-flow-state and we could "hitch a ride" or "ride the wave" along, maybe we can recover some of this unbrokenness of attentivity.

New comment by shomp in "For thirty years I programmed with Phish on, every day"

shomp — Sun, 03 May 2026 19:13:52 +0000

"This is engineering. I keep being told that" -- made me laugh out loud. We've gone from painting, to pointing to finished pieces in catalogues. Not to be utterly unhinged and tone-deaf but, there are other genres, such as jazz and screamo and noise that might suit the mood more aptly.

New comment by shomp in "Caveman: Why use many token when few token do trick"

shomp — Sun, 05 Apr 2026 15:09:01 +0000

everyone who thinks this is a costly or bad idea is looking past a very salient finding: code doesn't need much language. sure, other things might need lots of language, but code does not. code is already basically language, just a really weird one. we call them programming languages. they're not human languages. they're languages of the machine. condensing the human-language---machine-language interface, good.

if goal make code, few word better. if goal make insight, more word better. depend on task. machine linear, mind not. consider LLM "thinking" is just edge-weights. if can set edge-weights into same setting with fewer tokens, you are winning.

New comment by shomp in "A Claude Code skill that makes Claude talk like a caveman, cutting token use"

shomp — Sun, 05 Apr 2026 15:00:22 +0000

me disagree

New comment by shomp in "LLMs learn what programmers create, not how programmers work"

shomp — Tue, 24 Mar 2026 00:05:27 +0000

Great observation. The brain of a programmer is still a "black box" to the feed-forward network of nodes . But in theory, if you pumped a lot of the live-coding videos from something like youtube into the process, you could get a bit of that "what's your approach"-erism to bleed into the model. There might not be enough material there to truly "train it to think" but it would be interesting to try and "fill the gaps" of black-box-ed-ness in the LLM with supplemental "here was the process that got us there" video feeds. The next natural move might actually be recording thousands of hours of footage of developers working with the LLMs directly like in Cursor or another IDE that has LLM live-pair-programming , maybe calling it "pair programming" is generous , but it might be a reasonable foray into teaching the next generation of LLMs the "thought process" behind things. In reality you'd be teaching it which files to inspect, which windows to open/close, which tools to switch to and focus on. And while it might be imperfect, it might just be enough.

New comment by shomp in "In vitro neurons learn and exhibit sentience when embodied in a game-world(2022)"

shomp — Sun, 08 Mar 2026 20:55:56 +0000

Describing the chemistry doesn't invalidate the metaphor. It's still the moment fertilization initiates embryonic development. Explaining the gears of a watch doesn't make "the moment it starts keeping time" any less real, nor does it explain time.

New comment by shomp in "In vitro neurons learn and exhibit sentience when embodied in a game-world(2022)"

shomp — Sun, 08 Mar 2026 19:23:10 +0000

The fact that the perceptron is modeled after the neuron should make this an unsurprising find, but the question of active sentience starting as early as in vitro should give everyone pause for where personhood begins. That zinc spark.

New comment by shomp in "Happy Zelda's 40th first LLM running on N64 hardware (4MB RAM, 93MHz)"

shomp — Sat, 21 Feb 2026 23:19:10 +0000

Cool, is there maybe a video demonstrating this?

New comment by shomp in "U.S. jobs disappear at fastest January pace since great recession"

shomp — Mon, 09 Feb 2026 23:32:32 +0000

If you follow the American school of economics (Henry C. Carey), tariffs are actually a good thing, mainly because: 1) other nations all have tariffs (against the US and other nations) making free trade a delightful delusional idea 2) tariffs protect lesser industrialized nations from refining/enhancing raw materials into more expensive goods and selling them back to them. The systematic offshoring of industrial potential to cheaper labor places basically un-industrialized the US. I think it's very short-sighted to say tariffs are bad. What was bad was the de-industrialization of the leading superpower. The cure, if we may call it that, might be bitter medicine. Bitter, but necessary.