"ASUS emailed us last week (...) and asked if they could fly out to our office this week to meet with us about the issues and speak "openly." We told them we'd be down for it but that we'd have to record the conversation. They did say they wanted to speak openly, after all. They haven't replied to us for 5 days. So... ASUS had a chance to correct this. We were holding the video to afford that opportunity. But as soon as we said "sure, but we're filming it because we want a record of what's promised," we get silence."
Asrock (sub-brand of Asus but seemingly independent in the product and dev side) has been fine for me over the ~10 years I've bought their mobos. There was the thing a few months ago with X870 mobos that were apparently frying CPUs, but I think that was not sufficiently proven to be their fault?
That said, in their X670 / B650 they have the same setting as what this article is about, and it could be equally as broken on the software side as Asus's is, but I wouldn't know because I don't use Windows so I disabled it.
All the consumer brands are pozzed. My last build (i7-14700K) used an MSI board. Their secureboot is still broken. The BIOS setup is complete mess, and all the settings are reset after a BIOS update. I have to unplug and replug my USB keyboard after a poweroff, or it doesn't work. But I insisted on a board without RGB lights, and that limited the selection. Computers are over.
Just a few days ago people were talking about this on the kicad discord. A chinese team made an open hardware x86_64 motherboard and published it not too long ago. Then they were essentially wiped off the face of the planet.
That was the day I learned you literally cannot develop a computer motherboard without Intel's permission. Turns out Intel is no different than the likes of Nintendo.
This makes me angry, so can anyone think of a legitimate steelman of their position?
Expect my view is consistent with reality, though: they’re chasing profits and getting away with it, so why go on the record and look bad if they can ignore & spend that time on marketing.
Re your project: I'd expect at least the demo to not have an obvious flaw. The "lifestyle" version of your bag has a handle that is nearly twice as long as the "product" version.
This is a fair critique. While I am merely a "LLM wrapper", I should put the product's best foot forward and pay more attention to my showcase examples.
It is kinda a meme at this point, that there is no more "publicly available"... cough... training data. And while there have been massive breakthroughs in architecture, a lot of the progress of the last couple years has been ever more training for ever larger models.
So, at this point we either need (a) some previously "hidden" super-massive source of training data, or (b) another architectural breakthrough. Without either, this is a game of optimization, and the scaling curves are going to plateau really fast.
Again - I'd argue that the extraordinary success of LLMs, in a relatively short amount of time, using a fairly unsophisticated training approach, is strong evidence that coding models are going to get a lot better than they are right now. Will it definitely surpass every human? I don't know, but I wouldn't say we're lacking extraordinary evidence for that claim either.
The way you've framed it seems like the only evidence you will accept is after it's actually happened.
Well, predicting the future is always hard. But if someone claims some extraordinary future event is going to happen, you at least ask for their reasons for claiming so, don't you.
In my mind, at this point we either need (a) some previously "hidden" super-massive source of training data, or (b) another architectural breakthrough. Without either, this is a game of optimization, and the scaling curves are going to plateau really fast.
a) it hasn't even been a year since the last big breakthrough, the reasoning models like o3 only came out in September, and we don't know how far those will go yet. I'd wait a second before assuming the low-hanging fruit is done.
b) I think coding is a really good environment for agents / reinforcement learning. Rather than requiring a continual supply of new training data, we give the model coding tasks to execute (writing / maintaining / modifying) and then test its code for correctness. We could for example take the entire history of a code-base and just give the model its changing unit + integration tests to implement. My hunch (with no extraordinary evidence) is that this is how coding agents start to nail some of the higher-level abilities.
the "reasoning" models are already optimization, not a breakthrough.
They are not reasoning in any real sense, they are writing pages and pages of text before giving you the answer. This is not super-unlike the "ever bigger training data" method, just applied to output instead of input.
I'm not betting any money here - extrapolation is always hard. But just drawing a mental line from here that tapers to somewhere below one's own abilities - I'm not seeing a lot of justification for that either.
I so hate this. I have repeatedly seen PDFs containing nothing but a QR code and text like "not valid if printed" - this is truly silly. QR codes were created to form a bridge between the physical and the digital world, exactly so people can print them out. If you want it to be digital-only for some reason, use NFC or Bluetooth or whatever.
Citation needed
>Let's not personify and demonize vines
who is personifying and demonizing vines? again, citation needed
reply