Hacker News: what

New comment by what in "I think Anthropic and OpenAI have found product-market fit"

what — Thu, 28 May 2026 00:16:24 +0000

When you have to train a new model every few months to stay competitive, discounting that cost is rather dubious.

New comment by what in "The real cost of owning a home"

what — Wed, 27 May 2026 01:36:03 +0000

Eh, not really? If you have money to invest after rent, you’ll also have money to invest after mortgage payments.

New comment by what in "The Eternal Sloptember"

what — Mon, 25 May 2026 05:21:38 +0000

Yes, just throw 2kloc over the wall for some feature. Your coworkers must love you.

New comment by what in "Migrating from Go to Rust"

what — Mon, 25 May 2026 05:05:12 +0000

You’re supposed to bubble errors up to the level that can appropriately deal with them? You don’t need to log them each step of the way.

New comment by what in "AI is a technology not a product"

what — Mon, 18 May 2026 05:09:51 +0000

Except the LLM runs a bunch of searches and summarizes what it finds… what happens when they have no search engine to piggy back off of?

New comment by what in "Native all the way, until you need text"

what — Mon, 18 May 2026 04:37:34 +0000

> without looking like it’s been marked up with tags or formatting instructions.

From your own quote. Embedding html absolutely violates the spirit.

New comment by what in "Credit cards are vulnerable to brute force kind attacks"

what — Sat, 02 May 2026 04:25:23 +0000

$20 for food cart chicken strips is the real scam.

New comment by what in "Credit cards are vulnerable to brute force kind attacks"

what — Sat, 02 May 2026 02:41:07 +0000

Because people use credit cards for the rewards (cash, mileage, whatever) or because they don’t actually have the money now. They don’t want to pay for a unique number for every transaction (which doesn’t actually preserve privacy since most of the stuff you’re buying online needs a shipping address) nor do they want the money immediately pulled from their bank account.

New comment by what in "We need a federation of forges"

what — Wed, 29 Apr 2026 22:39:26 +0000

Don’t they already have extensive block lists that you can “subscribe” to? I think some official blue sky account was added to some and they got super mad?

New comment by what in "The AI industry is discovering that the public hates it"

what — Sun, 26 Apr 2026 00:48:43 +0000

They’ll just find a way to have $0 of profit. You have nothing to tax.

New comment by what in "Show HN: HNswered – watches for replies to your Hacker News posts and comments"

what — Sat, 25 Apr 2026 02:48:31 +0000

I just want to say please no. Comment chains where two people are arguing are the worst. Negative value. I’m always confused how they happen here, are people just refreshing their comments page waiting for a reply?

New comment by what in "The operating cost of adult and gambling startups"

what — Sat, 25 Apr 2026 02:07:33 +0000

You must not leave the house. Emergency services responding to ODs is commonplace in SF. It happened at least once per week outside my office. Walgreens (while they were still open) ran audio ads in the store encouraging you to buy narcan.

New comment by what in "Refuse to let your doctor record you"

what — Sat, 25 Apr 2026 01:41:15 +0000

> It sits next to the doctor helping them focus on you by transcribing the session, it doesn't do anything the doctor can't and definitely doesn't do anything the doctor SHOULD

You said the transcript isn’t available, only the notes/summary. The notes is what the doctor should do, the AI should only transcribe for the doctor’s review.

https://news.ycombinator.com/item?id=47895868

New comment by what in "Refuse to let your doctor record you"

what — Sat, 25 Apr 2026 01:33:29 +0000

You said you evaluate the error rate every month. How can you do that if you don’t have the recording or transcript?

New comment by what in "Our eighth generation TPUs: two chips for the agentic era"

what — Thu, 23 Apr 2026 03:16:46 +0000

It’s not about creativity. The incentive to produce drops to zero when an LLM is just going to slurp it up and regurgitate it without some form of compensation (notoriety, money, whatever).

New comment by what in "Our eighth generation TPUs: two chips for the agentic era"

what — Thu, 23 Apr 2026 01:20:05 +0000

Which ever shitty model they’re using for search is so much better than the free offerings from the other companies. It’s not even close. It’s not going anywhere.

New comment by what in "Drunk post: Things I've learned as a senior engineer (2021)"

what — Wed, 22 Apr 2026 04:50:44 +0000

And this will get you like $1M at 45? You can’t retire on that.

New comment by what in "ChatGPT Images 2.0"

what — Wed, 22 Apr 2026 04:19:38 +0000

Shouldn’t every model get the same prompt? Seems a bit weird, especially when you can’t see the prompts that were used.