Hacker News: ekms

New comment by ekms in "PhD Timeline"

ekms — Mon, 28 Apr 2025 11:50:29 +0000

I think the perception of America was just as critical, or even moreso, during the GWB presidencies and iraq war. (I imagine nixon and the vietnam war were thought of similarly). There's a tendency for people to have some historical amnesia and think of trump as qualitatively different somehow. But I can assure you that GWB was truly reviled especially in northern europe metro areas, and Americans was viewed then similarly to how it's viewed now.

New comment by ekms in "Startup Winter: Hacker News Lost Its Faith"

ekms — Tue, 21 Jan 2025 20:04:16 +0000

I stopped reading hn regularly a couple years ago just because the top-voted comments were so reliably negative cynical takes. And not just for articles related to startups.

New comment by ekms in "Feds arrest telehealth execs for overprescribing Adderall"

ekms — Mon, 17 Jun 2024 07:36:08 +0000

I guess they take it for the same reason you do. The effects aren't really different in people with ADHD vs people without [0]

> I cannot tell you how much literature there is trying to convince you that Adderall will not help healthy people, nor how consistently college students disprove every word of it every finals season. That makes “only give Adderall to people with ADHD” a moral judgment, not a medical one. Adderall doesn’t “cure” the “disease” of ADHD, at least not in the same way penicillin cures syphilis. Adderall will give everyone better concentration, and we’ve judged that it’s okay for people with terrible concentration to use it to overcome their handicap

[0] https://slatestarcodex.com/2017/12/28/adderall-risks-much-mo...

New comment by ekms in "Mindfulness interventions for teens decrease mindfulness, study finds"

ekms — Mon, 13 May 2024 04:12:13 +0000

I dont think the original series of guided meditations from the headspace founder was aimed at productivity. From what I remember, it's a pretty typical breath- and bodyscan-centered vipassana style. I can't comment on any of the subsequent instructors or lessons though, haven't tried them

New comment by ekms in "Leaving Rust gamedev after 3 years"

ekms — Fri, 26 Apr 2024 20:34:19 +0000

I mean, he does like a good rant lol. But this seems like a bad take. The witness came out ~8 years ago, and Braid came out ~8 years before that. Braid Anniversary is launching next week, he's actively developing his language and next game (occasionally streams). "he's just resting on his laurels now" I think is clearly wrong

New comment by ekms in "Humane AI – Pico Laser Projection – AI Twist on an Old Scam (2023)"

ekms — Wed, 17 Apr 2024 17:09:14 +0000

New comment by ekms in "Humane AI – Pico Laser Projection – AI Twist on an Old Scam (2023)"

ekms — Wed, 17 Apr 2024 14:00:51 +0000

I assume when people talk about Jobs bulshitting they’re more so referencing the earlier Lisa or NeXT eras

New comment by ekms in "Use Your Potions and Scrolls"

ekms — Wed, 17 Apr 2024 02:05:26 +0000

Did you watch any of lifecoach's slay the spire streams? He has some very impressive win rates/streaks and is I think much more conservative with potion use than most players.

New comment by ekms in "Lore Harp McGovern built a microcomputer empire from her suburban home"

ekms — Mon, 08 Apr 2024 20:21:45 +0000

23% > 20% which means if someone goes into the field of computer programming they're more likely to remain in the field if they are a woman than if they are a man. "remain in the field" is used as a proxy for success.

You could argue about whether or not it's a good proxy for success, but your response sounds like you think women would be more likely to drop out of the field alltogether than men, which doesnt appear to be true

Ask HN: Tips for Coding Effectively with Copilot?

ekms — Mon, 08 Apr 2024 19:24:53 +0000

I'm trying github copilot for the first time on a new project. So far I'm having trouble getting much useful output from it. Basically it just provides the types of suggestions a conventional autocomplete tool would (function signature completion, method suggestions) except with some random mistakes thrown in. Are you writing long instructive comments in your code in order to generate good output? Are there specific languages or frameworks it works better with than others?

Comments URL: https://news.ycombinator.com/item?id=39972845

Points: 1

# Comments: 0

New comment by ekms in "Sam Bankman-Fried sentenced to 25 years in prison"

ekms — Thu, 28 Mar 2024 21:28:36 +0000

Hm, not sure why people are downvoting this ^ Here's a source: https://www.wired.com/story/ftx-bankruptcy-bitcoin-value/

New comment by ekms in "Sam Bankman-Fried sentenced to 25 years in prison"

ekms — Thu, 28 Mar 2024 19:24:14 +0000

Did you get paid back yet? From what I read the ftx estate claimed to be able to make all claimants whole

New comment by ekms in "GPT-4V(ision) Unsuitable for Clinical Care and Education: An Evaluation"

ekms — Tue, 26 Mar 2024 21:02:15 +0000

Wish we'd get more articles from actual practitioners using generative AI to do things. Nearly all the articles you see on the subject are on the level of existential threats or press releases, or dunking on mistakes made by LLMs. I'd really rather hear a detailed writeup from professional people who used generative AI to accomplish something. The only such article I've run across in the wild is this one [0] from jetbrains. Anyway, if anyone has any article suggestions like this please share!

https://blog.jetbrains.com/blog/2023/10/16/ai-graphics-at-je...

New comment by ekms in "Sora: Creating video from text"

ekms — Fri, 16 Feb 2024 04:49:52 +0000

You do know that video wasnt like... state of the art video generation a year ago? It's an intentionally silly meme video

New comment by ekms in "Memory and new controls for ChatGPT"

ekms — Tue, 13 Feb 2024 21:47:54 +0000

Is it better at those types of things than copilot? Or even just conventional boilerplate IDE plugins?

New comment by ekms in "Vision Pro teardown part 2: What’s the display resolution?"

ekms — Wed, 07 Feb 2024 23:25:51 +0000

Some people have much worse vision? Makes sense that people would perceive visual clarity differently

New comment by ekms in "'Stupid,' 'shameful:' Tech workers on Y Combinator CEO Garry Tan's rant"

ekms — Fri, 02 Feb 2024 03:00:14 +0000

> implying that wasn't the case always on twitter?

New comment by ekms in "Apple Vision Pro review"

ekms — Thu, 01 Feb 2024 05:52:11 +0000

I think parent comment was maybe trolling a bit with a memoji comparison lol. Hard to say though

New comment by ekms in "Hasbro laying off Wizards of the Coast staff is baffling"

ekms — Sun, 17 Dec 2023 02:31:15 +0000

woah. it's been years since ive seen a penny arcade comic. didnt realize it was still around!

Show HN: Solving NYT Connections with ChatGPT

ekms — Tue, 05 Dec 2023 20:11:40 +0000

Just for fun I decided to see if I could use chatGPT to solve NYT Connections word puzzles. It uses a pretty straightforward BFS search in which the LLM is first prompted to generate several possible groupings of four related words, and then a different prompt is used to evaluate the soundness of each of those groupings.

This approach seems to be able to produce the correct solution somewhat less than half the time. Some observations:

* For whatever reason, chatGPT-4 seems to be a bit worse than 3.5 at generating Connections groupings. I haven’t tested systematically so maybe this is just some small sample size bias. But at the very least it isn’t obviously better

* It really struggles with the “words that can fill in the blank” style groups. Often it will correctly come up with the right category (e.g. “words that can precede `cheese`”) but will only be able to identify 2 of 4 words in that grouping

* It frequently generates very vague categories (“words that can be nouns”) despite nothing like that appearing in the proposal prompt. Also it will still sometimes score them highly, despite there being several explicitly examples in the value prompt disallowing these types of categories

If you have any idea for how to improve this, please let me know (or send a PR)!

Comments URL: https://news.ycombinator.com/item?id=38536319

Points: 8

# Comments: 0