HNHacker News
TopNewBestAskShowJobs

lupire

8,929 karma · joined October 17, 2018

submissionscomments
lupire··on All clocks are 30 seconds late
Countdown clocks run backwards, not forwards.
lupire··on All clocks are 30 seconds late
It depends on the application. Are you summing times (as with a pay clock at a job), or are you paying for error in both directions for some reason?
lupire··on All clocks are 30 seconds late
Technology Connections did it:

https://m.youtube.com/watch?v=NeopkvAP-ag

Bonus on analog vs digital mechanism in flip clocks:

https://m.youtube.com/watch?v=ZArBfxaPzD8

lupire··on All clocks are 30 seconds late
Quartz is an analog mechanism like a pendulum. It doesn't stop at each cycle.
lupire··on Uncut Currency
Wait till you find out what paintings cost and what paint coats.
lupire··on Uncut Currency
Per influencer rules, if you screws around neither someone's head for lolz, you should pay them combat wages for their time afterward.
lupire··on Uncut Currency
Steve Wozniak is white and soft looking.
lupire··on Uncut Currency
You don't have to. You could be a sane, intelligent person and not look at Instagram at all.
lupire··on Human study on AI spear phishing campaigns
Did you report phishing before you clicked? If not, you deserve to be dinged.
lupire··on Human study on AI spear phishing campaigns
You left your ID in the baggage?
lupire··on Human study on AI spear phishing campaigns
It's only understandable because no one has standards.

If I talk to Peter, Paul has no business getting any information about that or discussing it with me, untl Peter introduces me to Paul.

We teach our kindergarteners this rule!

lupire··on Human study on AI spear phishing campaigns
Al legit bank will NEVER legitimately call you, except to say "call us back at the number on your card" . Caller ID is not secure.
lupire··on Perplexity got ads
Yes, we know that individual consumers in general are not economically rational. The modern economy is built on this thesis.
lupire··on How saffron became an American cash crop
It pairs well with a $25 (plus tax and shipping) single, frozen, partially-cooked hamburger that you gave to finish preparing yourself.. https://www.goldbelly.com/restaurants/gramercy-tavern/the-gr...

They are rich retired people selling miniscule quantities to other rich people. It's just for fun. The excess cost of the saffron is lower than the transaction cost of fulfilling the order.

lupire··on What we learned copying all the best code assistants
Anton Ego: "Not everyone can cook, but great cook can come from anywhere"
lupire··on The Evolution of SRE at Google
ChatGPT is rarely available??
lupire··on 4.5M Suspected Fake Stars in GitHub
You are describing small communities, not collectivist ones.

Large "Collectivist" communities have body count in the hundreds of millions.

lupire··on iTerm2 critical security release
Better would say "After the immediate problem is patched, I will post more details about future plans for security, probably within X days. I welcome suggestions on the feature tracker."

Empty vague promises aren't really better than being quiet. They rely entirely on the reader's good faith, but if that good faith exists (which it likely does for this excellent product and excellent developer), then the promise adds no information.

lupire··on iTerm2 critical security release
You can expect anything you want in software you use, and choose not do you software that fails to meet expectations.

A software author who takes pains to publish his work and who accepts financial donations, is likely interested in maintaining his reputation and improving his skill and quality.

Finally, security bugs are in a class of their own. Giving out free junk is OK. Giving out free secret poison is not.

lupire··on iTerm2 critical security release
This is why all "temporary" settings go in .gitignored files or equivalent,

or on no-merge testing branches,

and we have test on the release branch that asserts for each config value that it is set to a safe value,

and we have a test that reflectively asserts that every config key has a value assertion test.

lupire··on There Are More Trees on Earth Than Stars in the Milky Way (2016)
(maybe, estimated)
lupire··on One of my papers got declined today
That's much less of a problem than the fact that papers are such poor media for sharing knowledge. They are published too slowly to be immediately useful versus just a quick chat, and simultaneously written in too rushed a way to comprehensively educate people on progress in the field.
lupire··on One of my papers got declined today
Gauss had plenty of room for slack, giving people time to catch up on his work..

Every night Gauss went to sleep, mathematics was held back a week.

lupire··on One of my papers got declined today
Andrew Wiles spent 6 years working on 1 paper, and then another year working on a minor follow-up.

https://en.m.wikipedia.org/wiki/Wiles%27s_proof_of_Fermat%27...

lupire··on Terence Tao: One of my papers got declined today
It's important to remember there a journal's reputation is built by the authors who publish there, and not vice versa.
lupire··on India's food shortage filled American libraries
It's a bit reversed. The history doesn't make people lazy, but lazy people rationalize their laziness by claiming clout from history.

Peoples who achieved more than their ancestors have less history to be prideful of, since they are proud of their present.

lupire··on 30% Drop In o1-Preview Accuracy When Putnam Problems Are Slightly Variated
They are not only simple renames. LLM is good at those. They are minor structural changes.
lupire··on 30% Drop In o1-Preview Accuracy When Putnam Problems Are Slightly Variated
The best human performance on that task required many many hours of private work given that input.

How much would ChatGPT charge for that much reasoning? Isn't cost quadratic in sort term working memory?

It would be more interesting to prompt it with X% of a new paper's logical argument, and see if it can predict the rest.

lupire··on 30% drop in O1-preview accuracy when Putnam problems are slightly variated
The researcher's answer to their variant of "Year: 2016 ID: A1" in the appendix is wrong.

The solution (sum of 1,2,5,6,9,10,13,14, ...) has an alternating pattern, so has to be two piecewise interleaved polynomials, which cannot be expressed as a single polyomial.

Their answer works for k=1,2, but not k=3.

https://openreview.net/pdf?id=YXnwlZe0yf

This does not give me confidence in the results of their paper.

lupire··on 30% drop in O1-preview accuracy when Putnam problems are slightly variated
When 4 came out they released a document that did BOTH inflate scores by changing the exam conditions, and also bragged about scoring worse than guessing on a multiple choice test.
← PreviousPage 17 of 34Next →