Hacker News: distrill

New comment by distrill in "Qwen 3.8 27B"

distrill — Fri, 14 Aug 2026 22:35:37 +0000

this sort of test surely only works when it's a brand new concept/prompt. like the exact same models trained newer data will surely get better at putting this together over time

New comment by distrill in "ChatGPT Is Down"

distrill — Wed, 15 Jul 2026 00:00:10 +0000

i also was able to continue using it in a web page that had already loaded, but can't load a new one

New comment by distrill in "ChatGPT Is Down"

distrill — Tue, 14 Jul 2026 23:50:00 +0000

upstream connect error or disconnect/reset before headers. reset reason: connection termination

New comment by distrill in "ChatGPT Is Down"

distrill — Tue, 14 Jul 2026 23:48:36 +0000

upstream connect error or disconnect/reset before headers. retried and the latest reset reason: remote connection failure

ChatGPT Is Down

distrill — Tue, 14 Jul 2026 23:48:36 +0000

Article URL: https://chatgpt.com/

Comments URL: https://news.ycombinator.com/item?id=48914446

Points: 13

# Comments: 11

New comment by distrill in "We might all be AI engineers now"

distrill — Fri, 06 Mar 2026 22:52:40 +0000

you make it seem like ai hesitation is a misunderstood fringe position, but it's not. i don't think anyone is confused about why some people are uninterested in ai tooling, but we do think you're wrong and the defensive posturing lines in the sand come off as incredibly uncurious.

New comment by distrill in "GPT-5.4"

distrill — Fri, 06 Mar 2026 22:29:36 +0000

this isn't really my opinion, but i think it's a perceived matter of _some_ principle vs just none, a lesser of 2 evils framing. if anthropic is on board with 99% of a government that i oppose, that could be seen as marginally better than openai being on board with 100% of a government that i oppose.

it does get a little weird thinking too hard about how the deal openai accepted was basically the same as the one anthropic was proposing. but this is my read of most of the sentiment in this direction.

New comment by distrill in "OpenAI raises $110B on $730B pre-money valuation"

distrill — Fri, 27 Feb 2026 20:23:28 +0000

i find codex 5.3 roughly on par with (though tbh still not quite as capable as) sonnet models, which are not even anthropic's flagship model family.

New comment by distrill in "Apple restricts Pebble from being awesome with iPhones"

distrill — Tue, 18 Mar 2025 22:01:57 +0000

i had the same problem with this behavior from google as i do from apple. i would be just as critical of google zealots blaming apple for google shortcomings as i am for apple zealots blaming google for apple shortcomings.

this is definitely an apple culture thing though. it's such a clear product choice to get apple users to pressure their friends into buying apple products.

New comment by distrill in "Apple restricts Pebble from being awesome with iPhones"

distrill — Tue, 18 Mar 2025 21:55:18 +0000

perhaps i'm out of date! this may have resolved with the recent increased support of rcs and i maybe haven't heard this complaint lately, it's worth checking into again.

New comment by distrill in "Apple restricts Pebble from being awesome with iPhones"

distrill — Tue, 18 Mar 2025 18:01:11 +0000

the thing i hear most frequently is naming group chats with mixed device users

New comment by distrill in "Apple restricts Pebble from being awesome with iPhones"

distrill — Tue, 18 Mar 2025 17:34:36 +0000

it's pretty frustrating how "apple people" just don't care that it's apples fault. i routinely hear my wife mutter "i hate google so much!" when a google maps integration is being intentionally hobbled to keep her using apple maps. or when she has trouble managing rcs conversations because somebody in our social world has the gall to be on an android phone.

New comment by distrill in "Two Grand Canyon-size valleys on far side of the moon formed within 10 minutes"

distrill — Mon, 10 Feb 2025 19:30:43 +0000

I thought it was quite good, although maybe it should have been a series rather than a single book. That might have given Stephenson time to flesh out the end, which felt a bit rushed.

New comment by distrill in "Show HN: Outerbase Studio – Open-Source Database GUI"

distrill — Wed, 04 Dec 2024 22:19:26 +0000

it runs in the browser

New comment by distrill in "Good programmers worry about data structures and their relationships"

distrill — Fri, 16 Aug 2024 19:11:00 +0000

Yes, I find myself in type hell with some regularity. TBH it happens with my own codebase too when libraries I want to use are authored by these type astronauts.

New comment by distrill in "Reddit down 502 Bad Gateway"

distrill — Thu, 25 Apr 2024 16:51:08 +0000

it just seems to be old.reddit, maybe this is how they're deprecating it

New comment by distrill in "The unwritten rules of negotiating with Meta"

distrill — Tue, 27 Feb 2024 19:54:00 +0000

from an eng pov, of course it is.

from any rational business pov, of course it's not.

New comment by distrill in "What can I do as an amateur hacker to get the best programming job next year?"

distrill — Mon, 28 Aug 2023 19:49:31 +0000

For a first job in the industry just spam applications. Assuming an acceptable resume that describes technologies you're competent with, and assuming it contains some projects or interesting things that have been built / worked on, some company somewhere will eventually take a flier on you.

Getting interviews after the first job I think is pretty self explanatory. Still spam applications but at that point you've already done this before.

FWIW, when I'm actively looking for work I will apply to maybe 10 or 15 places a day, often with cover letters that sound fairly bespoke (it is pretty easy to customize a generic template in a way that doesn't sound forced). It is not uncommon for me to have applied to 100 places by the time I'm busy enough with interviews to be comfortable stopping the application process.

New comment by distrill in "What can I do as an amateur hacker to get the best programming job next year?"

distrill — Mon, 28 Aug 2023 19:40:52 +0000

It is pretty universally used to filter out candidates. FAANG pioneered this and maybe for a while you could work in startups without grinding programming challenges, but unfortunately today even startups lean pretty heavily into this.

New comment by distrill in "What can I do as an amateur hacker to get the best programming job next year?"

distrill — Mon, 28 Aug 2023 16:40:40 +0000

Honestly, it's a chore and we all know it sucks but grinding leetcode will open up the most doors.