HNHacker News
TopNewBestAskShowJobs

l33t7332273

760 karma · joined August 8, 2023

submissionscomments
l33t7332273··on Big advance on simple-sounding math problem was a century in the making
We know the primes are rich in arithmetic progressions (and in fact, any set with positive “upper density” in the primes is also rich in arithmetic progressions).

So we do know that there are 100,000,000,000! primes that are equidistant from one another, which is neat.

l33t7332273··on Big advance on simple-sounding math problem was a century in the making
Less than 100 years later, we stand waiting for nuclear bombs guided by GPS to be launched when the authorization cryptographic certificate is verified.
l33t7332273··on Counterintuitive Properties of High Dimensional Space (2018)
Luckily if you do enough math it becomes much easier to reason about such spaces
l33t7332273··on Show HN: I built Bazaari.io to make launching online stores simple and fast
Amazon’s is also a smile.

I think your logo looks good, but the last thing a new business needs is the suits coming in and shutting them down

l33t7332273··on Heart rate variability differs between smartwatch and clinical testing
> Researchers found the method Apple uses, standard deviation of normal-to-normal, or SDNN, is the most accurate. Other brands implement a system called root mean square standard deviation, RMSSD, which Tenan said produces a wider range of error in the measurement

This is interesting for a couple of reasons. Firstly, I have one so it’s cool to know Apple designed it well in that respect, and that I shouldn’t lose confidence in the readings I get.

The second reason it’s interesting is that this kind of decision, which time-domain measure of heart rate variability metric to use, is the kind of super small design decision that really makes a system special.

Having really good developers and designers really shines through for moments like these.

l33t7332273··on The phone ban has had a big impact on school work
They’re not even talking much of the time, they’re texting (or worse, scrolling some social media)
l33t7332273··on The phone ban has had a big impact on school work
This is why I think managing notifications is so important.
l33t7332273··on Executives and Research Disagree About Hybrid Work. Why?
The research in this area has been hazy at best. It’s very difficult to quantify all of the factors.

It’s a disservice to pretend that WFH > RTO is “settled science”

l33t7332273··on Understanding the Limitations of Mathematical Reasoning in LLMs
I thought attention was all you need
l33t7332273··on How private intelligence companies became the new spymasters
I think it’s GP’s own evaluation
l33t7332273··on It's Time to Stop Taking Sam Altman at His Word
Just wondering if you’ve actually hidden inventions you would actually consider to be great?
l33t7332273··on X capitulates to Brazil's Supreme Court
Can you elaborate who “we” is and why you think more such speech js good?
l33t7332273··on X capitulates to Brazil's Supreme Court
Do you think there is any speech that can be dangerous to democracy?

Do you think such speech should be allowed to exist?

l33t7332273··on Scaling up linear programming with PDLP
I second this blog. The comparison of (10 different!) MIP modeling techniques for piecewise linear functions is more complete than any I’ve seen.
l33t7332273··on The centrality of stupidity in mathematics
There are lots of areas of math grad school that are well understood. Pretty much everything up to quals (and some beyond that) is well known and teaching methods are far from absent.
l33t7332273··on Terence Tao on O1
I was referring to section 4 of A Survey for Solving Mixed Integer Programming via Machine Learning(2024): https://arxiv.org/pdf/2401.03244.

I’ve heard (but not so much observed) that there is substantial difference between recent models, so it’s possible that they are better than when this was written.

Anyways, CPLEX has an associated modeling language that features syntactic sugar which has the effect of providing opaqueness to the underlying MILP that it solves. I find LLMs essentially unable to even make an attempt at determining the MILP from that language.

PS: How is Xpress? Is there some reason to prefer it to Gurobi or Mosek?

l33t7332273··on Terence Tao on O1
I an also working in OR and I have had the complete opposite experience with respect to MILP optimization(and the research actually agrees; there was a big survey paper published earlier this year showing LLMs were mostly correct on textbook problems but got more and more useless as complexity and novelty increased.)

The results are boiler plate at best, but misleading and insidious at worst, especially when you get into detailed tasks. Ever try to ask a LLM what a specific constraint does or worse ask it to explain the mathematical model of some proprietary CPLEX syntactic sugar? It hallucinates the math, the syntax, the explanation, everything.

l33t7332273··on Keyhole – Forge own Windows Store licenses
There was that Windows IPv6 no click zero day within the last couple of weeks
l33t7332273··on Achieving high Python performance with code generation (2022)
Some things aren’t feasible to use another language for, but I see your point.

My use case is MILP optimization. Constructing the model representation(the step necessary before calling some optimizer binary) is very slow in CPython, and Pypy makes development much less painful, since you often need to construct many models to get it right.

l33t7332273··on Achieving high Python performance with code generation (2022)
I’ve just started using pypy and it is night and day better. It’s truly incredible.
l33t7332273··on Physics is unreasonably good at creating new math
I’m struggling to see your point. My interpretation is that in each of those times, we used some math to talk about things we couldn’t observe until we were able to make experiments to observe them.

Are you saying that one day we will be able to devise experiments to observe these things?

l33t7332273··on Athletes and musicians pursue virtuosity in fundamental skills
I disagree. “Results” can pretty hard to quantify in athletics, especially in team athletics. Sure, there is a binary result for a win or loss, but are so many confounders that it can be pretty difficult to determine if a given strategy, diet, or gear is optimal, or even better than another.
l33t7332273··on Consistently Making Wrong Decisions Whilst Writing Recreational C
My intro to systems course did something similar with shared memory calls if I recall correctly
l33t7332273··on Go is my hammer, and everything is a nail
My Go is a little rusty by now, but I thought they supported some type of dynamic linking(although if I recall correctly it comes with a number of free footguns)
l33t7332273··on Study shows that tacking the “AI” label on products may drive people away
Many people who do know what LLM stands for (and could implement the architecture themselves) are also skeptical about why anything today is going to be better than the last 20 years of snake oil “AI”.
l33t7332273··on 0xCAFEBABE & 0xFEEDFACE (2003)
What objectionable thing is 0x73571C23 supposed to be?
l33t7332273··on GPT-4o System Card
I’m trying to come up with an intuitive reason the model would do that and I’m seeing nothing.

Deep neural networks are very strange beasts.

l33t7332273··on Mapping Hacker News to find who knows what in the HN community
I think that especially in CS, since applications of which touch on nearly every possible field of knowledge, computer scientists often run into trouble of assuming they know more than we do.
l33t7332273··on CrowdStrike Update: Windows Bluescreen and Boot Loops
We ran our code on our own accounts on the school’s system.
l33t7332273··on CrowdStrike Update: Windows Bluescreen and Boot Loops
I think this is more of a failure on the software development side than the domain specific functionality side.
← PreviousPage 3 of 13Next →