HNHacker News
TopNewBestAskShowJobs

HarHarVeryFunny

8,342 karma · joined July 29, 2017

submissionscomments
HarHarVeryFunny··on OpenAI is well positioned to fast-follow Jev
Jack of all trades, master of none.
HarHarVeryFunny··on OpenAI is well positioned to fast-follow Jev
A lot of business automation doesn't need AGI, doesn't want to pay for AGI if they don't have to, and would be better off using a classifier than something increasingly unreliable with a mind of it's own.

There are many automation pipelines that use LLMs because there was no choice, but the multi-way classification that Jev provides is exactly what they need, and is going to be way faster and cheaper, as well as having the benefit of calibrated probabilities and structured output that can be relied on.

HarHarVeryFunny··on OpenAI is well positioned to fast-follow Jev
Jev seems to have three benefits:

1) It's very cheap and fast - you provide one input and many potential classifications, and the compute to ingest the input is shared.

2) It generates structured output natively - guaranteed to be correct

3) It's output probabilities are calibrated to actually mean something

OpenAI, or anyone else, could certainly replicate it - there are already articles guessing how Jev achieves its "parallel" classifications, but it seems the AI companies need to decide are they in the business of providing intelligence/tokens, or are they in the application business trying to compete with all their customers (not that Jev uses OpenAI).

HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
It's obviously for PR, but has rather backfired.

OpenAI don't seem to have much direction. First they seemed to be concentrating on consumer with ChatGPT, while also pursuing things like image and video generation, about to get into porn (sex-chat), then they suddenly woke up to Anthropic owning the much more lucrative business/API and coding markets and dropped those distractions to compete with Anthropic instead. The whole Navier-Stokes PR debacle was also due to focusing on Anthropic (who they had heard rumor had solved it), maybe also in part jealous of DeepMind's positive AI-for-science press.

So, is this (poorly conceived) PR a rational attempt to boost investment, or more of a case of paying too much attention to the competition?

HarHarVeryFunny··on Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM
I was indeed channeling Apple - presumably most people on a tech forum realize that.
HarHarVeryFunny··on NASA’s Mars Sample Return mission is dead
Yep - the Chinese space program seems very long term and strategic in nature, both incrementally working towards moon landing and moon base, and beyond. It's obviously great for inspiring generations of kids to get into STEM, as well as national prestige, maybe a good driver for things like robotics and AI as well.
HarHarVeryFunny··on NASA’s Mars Sample Return mission is dead
Guess we'll have to wait for China to do it and hopefully publish the results.

They are scheduled to go in 2028.

https://en.wikipedia.org/wiki/Tianwen-3

HarHarVeryFunny··on NASA’s Mars Sample Return mission is dead
Yeah, and his mom was just in the news saying he really wants to live on Mars, so he can tweet about the samples in-between fertilizing his potatoes. :)
HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
In the world of PR it's about appearances - who grabs the headlines.

It this had happened and been shared with mathematicians, there would be leaks.

HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
I wouldn't expect the person who paid the $15M bill to be the one to take the bet, but the "expect to solve/announce it soon" rumor apparently came from some OpenAI employee, so the ship seems to be leaking, and yet nothing more.

I can see Altman sitting on 100 less famous math results waiting for the best PR opportunity to release them, but hard to see him sitting on this, risking being scooped by a competitor.

HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
I guess the OpenAI board was an advisory group too.

  Importantly, our board will not be responsible for advising us on anything.
HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
Who knows if they are behind or not?! Things like advanced math and advanced hacking require you to put a ton of effort into specifically training for those skills, and I've not seen any indication that the Chinese AI companies are interested in doing that - they seem more business focused.
HarHarVeryFunny··on Advisory Group on Mathematics and Artificial Intelligence
> Rumor was they were sitting on the solution to the Hodge conjecture

I think someone at OpenAI (Brockman?) claimed they had "made substantial progress on it", but there is a mathematician on Twitter offering a $25,000 bet, with no takers, that they have not solved it, and that the "substantial progress" they have probably made would still leave them a lot further from a solution than they imagine.

I also see someone else on Twitter claiming to have used AI to understand what the Hodge conjecture is. :)

I really wish OpenAI would fire their PR department (incl. mathematicians), and concentrate on trying to build AGI, or RSI, or generate tons of money, or whatever they are trying to do.

Edit: Apparently not Brockman - some unknown leaker saying there was progress or an expected result soon

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
You can theorize all you like, and many people like yourself are expecting to see GDP growth pick up as a result of AI, maybe if only because of the boom in datacenter construction etc (regardless of whether AI ends up actually boosting the economy), BUT ...

The reality is that so far there has been no sign of GDP growth picking up. It is basically flat at 2.5% +/- over the last few years.

In a similar vein one might have expected that the internet (think of all the e-commerce and efficiencies!) might have shifted GDP growth into a higher gear, but it did not, although in that case there was at least a significant boost in the 1996-2000 "dot com" era when the build out was happening (then to be followed by the crash and all the unused dark fiber etc).

So, maybe the hoped-for AI boom will be just as much of a dud (as it appears to be so far) as the internet boom. New day, different tools, same growth.

It's perhaps odd that we're not even seeing datacenter/etc build out register on GDP, but perhaps the scale of it is not as large as the internet build out?

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
Huh?

Developed countries tend to have lower fertility rates (i.e. less kids per family) than undeveloped ones. Comfortably off, well educated people tend to put thought into whether to have kids or not. In poor countries people tend to have large families so they can send the kids out to work.

Having a high fertility rate isn't "high status" and a sign of a country's progress, but rather the exact opposite.

Of course this is separate from this issue of having a shrinking population, which is a sign that your citizens are not optimistic about the future, not a positive thing.

HarHarVeryFunny··on Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM
Mini-AGI is a totally inappropriate name - it seems what this project is shooting for, but not delivering on, is being a language model with "continual learning".

Where it seems to fail, by design, on this goal is in delivering continual learning that is more than just "memorization with LRU catastrophic forgetting".

That said, props to the author for thinking different and actually implementing something. Maybe the project can grow into something more, or inspire different ideas, if they continue to work on it.

HarHarVeryFunny··on Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM
In fact it says the opposite - that there is pruning.

Our brain also has some capacity limit, and maybe degraded memory performance over time, but in either case it's a graceful degradation - you may forget fine details of things that happened a long time ago etc, but you don't forget how to ride a bike just because it's been a while.

Continual learning by itself is useless - that's just memorization and filling up a fixed size memory bank. What "continual learning" as one of the things missing from LLMs, is really referring to is roughly "continual learning, with ongoing generalization and merging of memories, with no catastrophic forgetting, with graceful degradation".

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
Don't worry, we've got Elon Musk trying to single-handedly repopulate America.

And if that fails, we've got "Idiocracy" to save the human species as the dumb-asses pump out babies as fast as you like.

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
It's hard to say that it's stupidity (at least not country-specific stupidity) - it seems that globally, presumably due to the state of the world along many axis, many people are choosing to have fewer children, and it's happening on a large enough scale to have major consequences for things like the size of your tax base, trying to keep your economy growing, demographic concerns over an aging population with no-one to look after or pay for them etc, etc.

Different countries are handling this in different ways, with encouraging immigration being the major one. Japan is facing a major lack of workers, with many businesses having to shut down as a result, but have decided to tackle it in a different way relying on advanced robotics as an alternative to immigration.

There are very few developed countries NOT having this problem. Most of the population growth is in poorer parts of the world like Africa.

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
What makes you think they don't have a domestic market?

The Chinese domestic EV market is the largest in the world.

Chinese factories are full of Chinese robotics.

Companies like Huawei are struggling to meet the domestic demand for AI accelerators, while growing volume very fast year over year.

The Chinese domestic AI market is massive, but we just don't hear much about it outside of the models that are also sold here. Can you even name the models that ByteDance (China's Meta) makes?

As far as exports go, China is a low cost producer for many things. If the US doesn't want to buy high quality EV's costing 1/2 the price of a Tesla, then there are plenty of other countries happy to do so. The US is only 5% of the global population - there are lots of other consumers out there.

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
It depends on what you are trying to make, but there are workarounds for most things, and it becomes more of a cost and volume issue than a showstopper.

If your yield is low and you still want the volume, then run more wafers, but then you need access to more DUV machines and more wafers. If your defect density is too high, then design smaller chips with a higher chance of avoiding defects, which is what Hauwei are doing - split processor into multiple chiplets (NVidia do this too, to raise yields and reduce cost).

HBM3E usually has 5-6000 vias, but doing this with multi-pattern DUV without defects is tough, so CXMT currently drop that to 3000, at a cost of some loss of thermal and voltage stability.

HBM:processor production ratios aren't what Huawei would like, so they mitigate this at system level by putting an optical memory bus on the GPU chip and sharing memory across the system.

Not everything is a huge LLM - smaller models like recommendation systems don't have so many parameters and need so much memory, so why waste HBM on them? ByteDance use their SeedChip acelerator for this, currently made by TSMC (using an older sanctions-approved 28mm process), which instead etches dense RRAM on die beside the processor.

There is also a time dynamic to this, with China still using pre-sanctions equipment and chips as their domestic alternatives ramp up to replace them. One interesting part of this is HBM ... HBM is very demanding to make, and everyone had problems with it, with initially only SK Hynix being successful. Samsung and Micron took a year or so to catch up, and during this time Samsung had made a ton of HBM that didn't meet NVidia's specifications, so ended up, pre-sanctions, selling it to China, where it has acted as a stockpile to carry them over as CXMT's domestic capacity ramps up.

China seems to be doing fine. Of course they would like sanctions lifted, moreso for HBM than anything else, but all that sanctions have really achieved is accelerating their semiconductor independence. They are still building 1T+ SOTA models, standing up 100K clusters of domestic AI accelerators, etc, etc.

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
On a related note, I was reading yesterday that apparently the real bottleneck for Chinese production of AI accelerators is HBM production, not processors or ASML equipment.

The lack of ASML EUV machines certainly hurts, and pushing DUV so hard results in abysmal yields of good chips, but you can compensate by running more wafers or making smaller chips, and the net result is that Huawei's Ascend production volume is limited by CXMT's HBM capacity not processor dies.

The problem is that HBM manufacture requires many steps (die thinning, via drilling, plating, alignment) where the equipment used by everyone else (Samsung, SK Hynix, Micron) is also blocked by sanctions, so the Chinese are having to develop all of this themselves too, which they have, but yields are currently low, even when using shorter HBM stacks.

HarHarVeryFunny··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
Why would you want/need to?

The advantage of HBM over regular non-stacked DRAM is memory bandwidth, which also requires a super-wide memory bus - 2048 bits wide for HBM4. Compare that to the 128 bit wide bus of a modern CPU.

So to take advantage of it on the desktop, or anywhere else, you need that 2048 bit wide bus, and a processor capable of consuming 2-3 TB of data per second!

These are not normal requirements, other than for a GPU.

HarHarVeryFunny··on The Millennium Problems for Biology
There may be a few scientists rejecting use of AI for that reason, but I would expect them to be a tiny minority, just as I would expect the number of developers who accept that AI is a powerful coding tool, but refuse to use it just because they want to feel superior, is very small!

A much more practical reason we're not yet seeing a lot of headline scientific mathematical breakthroughs is just that it is ungodly expensive! e.g. The Navier-Stokes result cost around $20M at API prices, and academics just don't have that kind of money to spend. You'll see more mathematical and scientific results from practitioners when either the cost of compute needed for these sort of brute force results is more in line with the size of academic grants, and/or the AI companies donate more compute to the scientific community.

There are different reactions from different mathematicians of course - Terrance Tao vs Cedric Villani, and no doubt a lot of shock at the speed of advance, but it seems the reasoned complaint why they don't want the AI companies themselves working on these problems is because the outcome is not the same - you get a result that in of itself may have been suspected or useless (Navier Stokes), but no write up of any new math or insights that were developed along the way, which is the real reason mathematics and people like Erdos pushed these famous problems in the first place - because they were expected to yield interesting mathematics, just as years of work on FLT had done. Imagine if instead of Wiles's work, and all that had gone before him, all we had was a $20M compute bill, hundreds of pages of impenetrable math, and a billion lines of Lean proving it was true?!

HarHarVeryFunny··on The Millennium Problems for Biology
Sure - the AI companies want some quick trophy kills to feature in their IPO prospectus, but they are not going to themselves be cracking the genuinely tough problems.

There is a difference between what's easy/hard for a human vs computer, and LLMs haven't changed that. You might expect a computer to be good at tasks requiring prodigious memory and compute, and it turns out that some of these long-standing math problems are of that nature - not requiring new breakthroughs but rather just massive exploration of what is already known and what they were trained on.

There will no doubt be more math results like this, but presumably also ones that are "hard for a human, easy for a computer", requiring massive search (e.g. find an example/counter-example cf Navier-Stokes & Jacobian conjecture) rather than creativity.

HarHarVeryFunny··on The Millennium Problems for Biology
It seems that physics -> chemistry -> biology is a layering of abstractions and/or emergent phenomena, and any explanations of mechanics at these higher levels is more along the lines of "how was this built from these lego blocks" rather than being some more fundamental discovery.

You keep on making assertions - perhaps you could back it up with some explanation of why you regard this as a more fundamental discovery (or however you would like to characterize it)?

HarHarVeryFunny··on The Millennium Problems for Biology
What aspect of that is an unsolved problem, and one that feels more fundamental rather than engineering / in need of explanation ?
HarHarVeryFunny··on The Millennium Problems for Biology
What seems ridiculous to me is putting together any such lists, even in math, and expecting/hoping the AI companies to solve them. The people working for the AI companies are not scientists or experts in anything outside of building LLMs. The best chance of making progress on genuinely tough (not just computationally challenging) tasks is to put advanced tools in the hands of actual scientists.
HarHarVeryFunny··on The Millennium Problems for Biology
What would be your idea of a biology problem that isn't an "engineering" one?

It seems that understanding biology could be characterized as trying to figure out "how did nature engineer this".

For that matter isn't all of science this way?

HarHarVeryFunny··on The Millennium Problems for Biology
The goal of the suggested "origins of life" problem, as written, isn't to guess at history - it's to create some similar plausible self-emergent system in the lab.

Seems a bit of an ambitious goal for a language model though! There are presumably dozens of steps before you get to an RNA-based full-blown cell, and it's only been very recently that humans have managed to make an artificial self-replicating "cell" (container) of any type in the lab.

← PreviousPage 9 of 34Next →