Hacker News: tibbar

New comment by tibbar in "Footguns with Postgres “at time zone 'UTC'”"

tibbar — Mon, 28 Sep 2026 14:46:21 +0000

This summer, with the help of AI, I found an inconsistency in the way Postgres handles timestamp vs. timestamptz comparisons under a DST spring-forward gap for the datetime_ops btree family [0]. Essentially there are scenarios where expression B > A and B < C, but also C = A, which can cause queries using a btree index (among other things) to return an incorrect result.

The assessment in the mailing list was that this was a bug, but there were no good ways to fix it.

> Backpatching a behavioral change like this seems awfully scary. For the moment I'm just contemplating what we could potentially change in master. So far I don't like any of the choices :-(

[0] https://www.postgresql.org/message-id/flat/CA%2BCOZaDmCuOds-...

New comment by tibbar in "Don't call yourself an artisanal programmer"

tibbar — Sun, 13 Sep 2026 05:42:52 +0000

I prefer the term "classically trained programmer." :-)

Letter from Dolly Parton to Eminem

tibbar — Wed, 26 Aug 2026 02:09:12 +0000

Article URL: https://twitter.com/Eminem/status/2092367272891310482

Comments URL: https://news.ycombinator.com/item?id=49443331

Points: 81

# Comments: 11

List of inventors killed by their own inventions

tibbar — Fri, 21 Aug 2026 22:37:25 +0000

Article URL: https://en.wikipedia.org/wiki/List_of_inventors_killed_by_their_own_invention

Comments URL: https://news.ycombinator.com/item?id=49394569

Points: 11

# Comments: 0

New comment by tibbar in "Being ambitious and being a dad"

tibbar — Wed, 19 Aug 2026 20:58:09 +0000

If 75% of people cannot find a good partner, that suggests that all involved would be well-served by learning how to be better partners themselves:)

New comment by tibbar in "Being ambitious and being a dad"

tibbar — Wed, 19 Aug 2026 20:47:11 +0000

I don't know, this feels kind of overly judgy. It's pretty healthy and normal to have some goals in life and even to try to do something impressive. The fact that he is trying hard to balance this with caring for his family seems pretty reasonable to me.

New comment by tibbar in "Being ambitious and being a dad"

tibbar — Wed, 19 Aug 2026 20:42:53 +0000

Interestingly, the case you linked was a purely civil action. However, in 2016, the DOJ announced that in the future, it would prosecute similar cases criminally. [0][1] Take that for what you will!

[0] https://www.justice.gov/archives/opa/pr/justice-department-a... [1] https://www.justice.gov/archives/opa/speech/acting-assistant...

New comment by tibbar in "Being ambitious and being a dad"

tibbar — Wed, 19 Aug 2026 20:30:47 +0000

I think that the drive and energy to build one's own business (even if the business itself is very banal, like a laundromat or app-builder) is impressive and counts as ambitious. It's not easy. I tried starting a business and couldn't hack it.

New comment by tibbar in "Ask HN: GitHub employees what's going on? Why?"

tibbar — Wed, 19 Aug 2026 04:07:54 +0000

This is nightmarish.

New comment by tibbar in "Being ambitious and being a dad"

tibbar — Wed, 19 Aug 2026 00:33:50 +0000

Unfortunately, I think that this is possible, but it requires a superhuman well of energy, sleep deficit tolerance, and benevolence that very very few people possess. :/

New comment by tibbar in "Ask HN: GitHub employees what's going on? Why?"

tibbar — Tue, 18 Aug 2026 20:49:09 +0000

The reality is that it's probably a skill issue. Scaling a mature platform 10x can be a really, really hard problem, they obviously don't have their arms around a solution, and are probably spending a majority of their time on ops to keep the bleeding down. In the old days you would have Jeff Dean come down from heaven and invent a new database for you or something like that. It doesn't really seem like Github has that kind of technical ability, so they're probably trying to cobble together Azure ops with internal bandaids while everyone internally is cranking out AI code and it's just not going to be enough.

Show HN: Quantum Foosball

tibbar — Tue, 11 Aug 2026 22:30:54 +0000

Article URL: https://polkerty.github.io/quantumfoosball/

Comments URL: https://news.ycombinator.com/item?id=49265380

Points: 2

# Comments: 0

Trystero

tibbar — Tue, 11 Aug 2026 20:00:06 +0000

Article URL: https://github.com/dmotz/trystero

Comments URL: https://news.ycombinator.com/item?id=49263642

Points: 2

# Comments: 0

New comment by tibbar in "Changes to Google DeepMind teams"

tibbar — Wed, 05 Aug 2026 18:22:18 +0000

The labs are all very interested in bio right now. Demis is working on Isomorphic rn (Google's version of that) but I could imagine a lab tempting him away to work on their equivalent if it had stronger momentum

New comment by tibbar in "Postmortem for Kernel Soundness Bug #14576"

tibbar — Sun, 02 Aug 2026 00:31:38 +0000

Reward hacking will continue to be a problem. For sufficiently astonishing AI-created results, we must remember to ask if AI has not instead found it easier to elaborately fool us.

New comment by tibbar in "Run Kimi K3 using 29 GB of RAM at 0.50 tok/s"

tibbar — Sat, 01 Aug 2026 04:49:17 +0000

effort level is a separate parameter, and is probably what they want

New comment by tibbar in "I flagged two research papers for fake authors and both were accepted as orals"

tibbar — Fri, 31 Jul 2026 03:28:43 +0000

I think Pangram is de facto measuring the default Claude voice. It doesn't fire for me on technical GPT writing.

Locked in An Escape Room with Jacob Tsimerman

tibbar — Thu, 23 Jul 2026 21:37:15 +0000

Article URL: https://www.begiant.ca/stories/people/jacob-tsimerman-math

Comments URL: https://news.ycombinator.com/item?id=49028389

Points: 3

# Comments: 0

New comment by tibbar in "[dead]"

tibbar — Sun, 19 Jul 2026 23:24:30 +0000

My entire office of nerdy engineers has been gathering 'round to watch the world cup for an entire month. Judging by this, the world cup does seem to be of some interest to hackers. ;)

New comment by tibbar in "Kimi K3, and what we can still learn from the pelican benchmark"

tibbar — Fri, 17 Jul 2026 19:48:15 +0000

Training and serving large models does require increasingly more compute, though. (The Chinese labs have clearly found some massive optimizations, but my point was that you'd think at some point even those optimizations wouldn't be enough to keep up with exponentially increasing model sizes.)