Hacker News: tshaddox

New comment by tshaddox in "Harness Engineering 101: The LLM is the brain, the harness is the body"

tshaddox — Mon, 19 Oct 2026 12:50:00 +0000

Surely the only problem isn’t the knowledge of how to train and the confidence it would work. Wouldn’t it also have required a 1 or 2 order of magnitude increase in computing power over 2005-era supercomputers to be able to train a modern LLM in a reasonable time (“reasonable” meaning “finish training before GPT-3 comes out in 15 years”)?

New comment by tshaddox in "Why is every rationalist blog post way too long?"

tshaddox — Thu, 15 Oct 2026 03:21:52 +0000

> Things survive for no good reason.

And sometimes the “good reason” that something survives for a long time is that the better alternative was difficult to invent/discover/implement. Which is obviously not a “good reason” to oppose the introduction of better alternatives.

New comment by tshaddox in "Why is every rationalist blog post way too long?"

tshaddox — Thu, 15 Oct 2026 03:19:48 +0000

I don’t think it’s reasonable. The age of an idea does not imply that it’s true, valuable, or good.

New comment by tshaddox in "Redesigning source control at Anthropic"

tshaddox — Tue, 13 Oct 2026 23:31:19 +0000

A student of history might note that power tools do indeed allow a worker to replace multiple workers not equipped with power tools.

New comment by tshaddox in "Redesigning source control at Anthropic"

tshaddox — Tue, 13 Oct 2026 20:22:54 +0000

I use the ChatGPT general purpose interface rather than Claude, and this is not at all my experience. ChatGPT will almost always do one or more web searches for any query about recent or historical events.

New comment by tshaddox in "Redesigning source control at Anthropic"

tshaddox — Tue, 13 Oct 2026 20:20:52 +0000

That’s kinda like asking why does any company make and sell any tool rather than just using the tool to make money? Why does Milwaukee sell power tools to contractors rather than keeping those tools to themselves to profit from contracting work?

New comment by tshaddox in "Postgres isn't slow, your storage is"

tshaddox — Mon, 12 Oct 2026 14:22:43 +0000

It’s probably cheaper per unit when they buy at massive scale, but that was also true before the hardware inflation of the past couple of years. So I’m sure they are still paying much higher prices than in previous years. Many cloud services have increased prices and I suspect that will continue for at least a few more years.

New comment by tshaddox in "The RAM shortage is bringing back DDR4"

tshaddox — Sun, 11 Oct 2026 20:03:17 +0000

> Econ 101 would be understanding that you can’t snap your fingers and have a new DRAM factory online overnight, and that this isn’t an indicator that the market is not functioning.

You can quibble over my usage of “functioning,” but all I’m referring to are the textbook characteristics of perfect competition. I’m not claiming that someone is to blame if any of these conditions don’t hold. Of course they never hold perfectly in practice!

I’m not trying to blame anyone for the fact that it takes a lot of money and time to get a new DRAM factory online. I’m just saying that this fact is bad for essentially everyone except for the small number of current DRAM producers.

> You’re defining a market “functioning well” as meaning the price is low for you and all demand is satisfied immediately without raising prices.

Well, yes, that’s pretty much the only reason to advocate for competitive markets, isn’t it? If you don’t want allocative efficiency and productive efficiency, then what do you want?

New comment by tshaddox in "The RAM shortage is bringing back DDR4"

tshaddox — Sun, 11 Oct 2026 17:13:04 +0000

The important question isn’t so much “are people trying to get the best deal possible for themselves?” but rather “how well is this market functioning?” Or phrased another way, “how competitive is this market?”

This particular market is pretty clearly not functioning well, and there are textbook Econ 101 reasons why a market may not function well. Of those textbook reasons, I suspect the most obviously applicable one here is large barriers to entry.

New comment by tshaddox in "Trump administration is suspending Microsoft from a green card program"

tshaddox — Thu, 08 Oct 2026 20:20:58 +0000

Not quite. That guy was appointed chancellor by the President after his political party won the most seats (still a minority) in two parliamentary elections that year. He had in fact lost the presidential election earlier that same year against the incumbent who ended up appointing him chancellor.

New comment by tshaddox in "Trump administration is suspending Microsoft from a green card program"

tshaddox — Thu, 08 Oct 2026 16:35:53 +0000

Several notable dictators came to power through elections or through the existing legal appointment process. Of course it's more common for a dictator to come to power through a revolution, military coup, or civil war, but the way a political figure originally came to hold their political position is not really relevant to determining whether they are a dictator.

New comment by tshaddox in "What is going on with ceiling fans"

tshaddox — Mon, 05 Oct 2026 01:23:46 +0000

I used to be a big metal fan.

New comment by tshaddox in "What is going on with ceiling fans"

tshaddox — Mon, 05 Oct 2026 01:22:50 +0000

This is most YouTubers who get very popular due to having original and opinionated content on niche topics. They tend to identify (perhaps correctly, perhaps not) their idiosyncratic personalities and opinions as the primary cause of their initial popularity and repeatedly double down on it as they grow in popularity. This is also the explicit advice that the vast majority of big YouTubers give to aspiring or up-and-coming YouTubers.

And I’m not advising that they should do anything different! Who am I?! I’m just pointing out that the culture of YouTubers is fundamentally about deliberately subjecting yourself to the most extreme audience capture.

New comment by tshaddox in "Big Tech ruined the cloud, so we're renaming ours"

tshaddox — Fri, 02 Oct 2026 16:22:57 +0000

I don’t know, this seems like a pretty basic feature that I would expect all the mainstream smart home ecosystems to support. Do they not? Is it possible with Alexa/Google Home/ etc. for anyone to say “turn off all the lights in the house” into any smart speaker? I’d honestly expect the defaults to only allow commands to apply to the devices in the same room/group as the smart speaker.

New comment by tshaddox in "PS5 Relapse Exploit"

tshaddox — Wed, 30 Sep 2026 16:07:38 +0000

That makes sense. But is the price difference between doing it on home hardware versus cloud hardware less significant than it is for LLM inference? And if so, is it mostly because LLMs need much more VRAM (or more exotic memory architectures)?

New comment by tshaddox in "PS5 Relapse Exploit"

tshaddox — Wed, 30 Sep 2026 05:02:13 +0000

I don’t know much about gaming GPU workloads and how they differ from LLM inference workloads, but I’m definitely rich enough to pay for an LLM subscription and definitely too poor to afford the GPUs it would take to get anywhere near that level of inference.

New comment by tshaddox in "Everybody’s home. No one’s coming over"

tshaddox — Wed, 30 Sep 2026 03:34:44 +0000

> It's worth noting that not everything that's good for you is enjoyable, especially in the moment.

And likewise, not everything you don't enjoy that society pressures you to do is somehow good and worth suffering through. Particularly when the whole point of the thing is ostensibly about enjoying yourself in the moment.

New comment by tshaddox in "Sonnet 5.5"

tshaddox — Tue, 29 Sep 2026 14:41:44 +0000

I suppose I am confused about why you’re confused. There’s a long history in computing of describing pieces of programming languages syntax syntax as “incantations” and similar. I suspect it has been very common, especially in the early part of developers’ careers, to know what you’re trying to do and to know that this code accomplishes it, but to not understand how the code works, to not be able to use the technique more generally, and to not understand all the effects your change has on the rest of the system.

New comment by tshaddox in "Sonnet 5.5"

tshaddox — Tue, 29 Sep 2026 02:04:17 +0000

I’m referring to competent engineers maintaining understanding over time of all the code they’ve produced. Long before agentic coding, codebases routinely grew beyond the comprehensive understanding of their own authors.

Of course less competent engineers (or anyone on a particularly disorganized or desperate day) can literally hand-write code they don’t understand even as they write it, but that’s not really what I’m talking about.

New comment by tshaddox in "Sonnet 5.5"

tshaddox — Mon, 28 Sep 2026 22:33:27 +0000

Humans could already produce more code than a human can understand. Even a single human in the pre-agentic era could produce more code than they could understand, certainly over a career and often even in the short term given the resources many companies give to maintenance.

A lot of old-school software engineering is about how to deal with this reality.