treprinum

I would suspect the physiological base of burnout is the depletion of mainly B vitamins in the brain. When you look at the Krebs cycle, the main energetic reaction in every cell, B vitamins play crucial role whether they act as catalysts of enzymes (B1, B2), provide electron transport (B2/B3 to FAD/NAD+) or are required by sub-processes of the cycle (B5, B7, B12). Being under constant stress is known to deplete B vitamins and at some point their lowered availability starts inhibiting higher cognitive tasks in favor of just survival. Diagnosing this is problematic as serum levels don't tell much about tissue levels.

Didn't Chinese scientists recently show a crazy success rate (~90%) of treating advanced Alzheimer/dementia by performing a microsurgery of the neck, allowing brain to dispose accumulated waste?

Nvidia just did what Intel/AMD should have done to threaten CUDA ecosystem - release a "cheap" 128GB local inference appliance/GPU. Well done Nvidia, and it looks bleak for any AI Intel/AMD efforts in the future.

Bad idea for social marketing - by the time a post becomes popular, it disappears.

When is Apple going to steal it and then sue Waterloo just to be sure?

Get admitted to online UIUC iMBA or BU OMBA, take the marketing/sales/entrepreneurship courses, quit.

That's a pity! CGTalk was the site where I first learned about Cg from Nvidia that later morphed into CUDA so unbeknownst to them, CGTalk was at the forefront of the AI by popularizing it.

Would that mean that we would need to exchange latent "embeddings" between various "reasoning" models for emulating thinking and an LLM will be just about converting to/from human language when interfacing with mere humans, at some point in the future?

I am really glad we are finding new pieces of the puzzle of how our gut works and perhaps can someday understand their effect on immunity, neuro-degeneration, cancer etc. for which we now only have accidental findings.

I said in some areas of Barcelona which is true. Now somebody did the Simpson's paradox using the whole data set instead of conditional data, leading to a different result.

So you cherry pick one real-world observation, try to refute it and it makes the rest wonderland? EU is doing super great because you just scentifically proved that Barcelona can't be as bad as LA?

No, but EU with all its social agenda is now imploding at unusual places that were once considered great places to live, inflation is doing its thing. In the US tech is considered the most important engine of the economy, EU wasted over a decade and lost its only competitive company that led one large sub-field of tech (Nokia). Pretending nothing is going on won't help the EU, sooner or later this will be apparent everywhere (with a chance a hot war with Russia collapses it quicker).

Why would anyone outside some desperate eastern European countries want to join? EU is going down hard, since 2008 no meaningful GDP growth, flat salaries, no tech companies propelling growth, old population, all the while US and Chinese GDP is exploding. The only thing EU has is regulation and high taxes, everything else is deteriorating quickly and the rate of decline is accelerating. Barcelona, previously a jewel of Europe, now has as many homeless as LA in some areas.

The top-end RDNA4 GPU will have 16GB RAM. That's a massive regression compared to 7900XTX and performance-wise it should be at best at the 7900XTX level. We are discussing AMD cards for LLM inference where VRAM is arguably the most important aspect of a GPU and AMD just threw in the towel for this cycle.

AMD decided not to release a high-end GPU this cycle so any investment into 7x00 or 6x00 is going to be wasted as Nvidia 5x00 is likely going to destroy any ROI from the older cards and AMD won't have an answer for at least two years, possibly never due to being non-existing in high-end consumer GPUs usable for compute.

Best way is to use ChatGPT + Whisper + ElevenLabs and a prompt where you tell the LLM to answer in German but to tell you what mistakes you made and what you should have said instead in your own language. Instead of this lazy "immersion" method German schools use in order to avoid speaking in any other language so students never understand what went wrong.

Tor browser asks you if you want to allow fingerprinting or not when a site attempts to query your HW info. Not sure why other browsers can't do the same.

I have an idea - how about Apple letting Mac users use Nvidia cards/chips again?

I remember somebody complaining about it on reddit, unable to overcome some BIOS limitation on an AMD G processor. Even on M3 Max one had to issue a special command to enable GPU to access more memory.

Quantized sure but there is some loss of variability of the output one can notice quickly with 30B models. If you want to use the fp16 version you are out of luck.

Problem with MI300x is the price. Problem with 7900XTX is that it's at best as good as Nvidia with the same RAM for a similar price. If 7900XTX had e.g. 64GB of RAM, was 2x slower than 4080, and kept its price, it would sell like crazy.

View it as Raspberry Pi for AI workloads. Initial stage is for enthusiasts that would develop the infra, figure out what is possible and spread the word. Then the next phase will be SME industry adoption, making it commercially interesting, while bypassing Nvidia completely. At some point it would live its own life and big players jump in. Classical disrupt strategy via low cost unique offerings.

AMD GPUs aren't very attractive to ML folks because they don't outshine Nvidia in any single aspect. Blasting lots of RAM onto a GPU would make it attractive immediately with lots of attention from devs occupied with more interesting things.

It's too slow and not very compatible. Most BIOSes also don't allow sharing that much memory with GPU (max like 16GB).

I ran some variation of llama.cpp that could handle large models by running portion of them on GPU and if too large, the rest on CPU and those were the results. Maybe I can dig it from some computer at home but it was almost like a year ago when I got M3 Max with 128GB RAM.

That's a question M3 Max with its internal GPU already answered. It's not like I didn't do any HPC or CUDA work in the past to be completely clueless about how GPUs work though I haven't created those libraries myself.

K80 used to be two glued K40 but their interconnect was barely faster than PCIe so it didn't have much benefit as one had to move stuff between two internal GPUs anyway.

I think plenty of enthusiastic open source devs would jump at it and fix their software if the software was reasonably open. The same effect as what happened when Meta released LLaMA.

For some reason Apple did it with M3/M4 Max likely by folks that are also on HN. The question is how many of the years spent designing HW were spent also by educating oneselves on the latest best ways to do it.

They are charging as much as Nvidia for it. Now imagine they offered such a card for $2k. Would that allow them to eat Nvidia's lunch?