Again, the problem is custom kernels in CUDA. It’s not straightforward for many applications (LLMs are probably the most straightforward).

It doesn’t sound like having a house will magically make you feel better. Plenty of people are just as exposed to markets as you yet respond differently. I suggest exercise, nutrition, nature, and extended travel if those are not already a part of your life.

Thank you. I sometimes do lose this perspective.

These kinds of comments make me think few people have actually tried. My experience has been 1 work day of getting things set up to work the same as before for training and testing (PyTorch).

You have to consider that the average person who tried to do machine learning on AMD GPUs got burned in the past decade and has no reason to change their opinion. Also in the past it was much harder to get access to cutting edge GPUs from AMD. The fact that AMD drops GPU support for ROCm quickly also earns them scorn. I don't think it is an unfair assessment. They earned their reputation.

ROCm has improved a lot. And you can rent mi300x in the cloud now. So if you have a program that runs on Nvidia GPUs, it takes no time to test it on a cloud mi300x. If it works you can use it and save some money in the process.

Unless you develop in CUDA, you can easily train code (e.g. PyTorch) written for training on Nvidia hardware on AMD hardware. You can even keep the .cuda() calls.

In theory. But if you actually work with that in practice, you're already going to have a bad experience installing the drivers. And it's all downhill from there.

I’m a AI Scientist and train a lot of models. Personally I think AMD is undervalued relative to Nvidia. No, chips aren’t as fast as Nvidia’s latest and yes, there are some hoops to get things working. But for most workloads in most industries (ignoring for the moment that AI is likely a poor use of capital), it will be much more cost effective and achieve about the same results.

Why would you want (versus need) a Tesla? It’s no longer an aspirational product, it’s an appliance.

For all you know he had a hard adjustment to the college workload. Maybe his high school was not serious and nobody really challenged him. You all could have interrogated why he may have struggled and shared your strategies for success, lifting him up rather than beating him down. That would make you true “brothers”.

What a weird “I hate fraternities” response. It’s clear you’d really struggle with a group that holds itself accountable.

My (biotech, mostly remote) company does this. They may not love it but they realize they have to hire from SF, Boston, NYC etc to get the best ML talent and people expect market salaries / don’t want to up and move if they don’t have to.

The issue is that actually, outside of technical paths, businesses do like to hire business grads as it allows them to do even less training and they usually don’t care if someone is well read or opinionated (possibly prefer the opposite).

Or, just maybe, AGI is a mirage with the bulk of its current utility as a marketing tool for much more realistic, if ultimately mundane, applications. OpenAI, of course, knows this.

Whether it's real or illusory, I feel like they turned the bitrate down significantly since launch. I'm not getting as many smart electrons per question as I used to.

This is measurably true. They claim they don't know why because "they haven't changed anything" but it is true at the very least

I think at least some of this is expectation setting. I'd been saving my prompts from the beginning. In some cases I saved the results.

On the API: My prompts are getting 'harder' and expecting more. My old prompts still work just fine still.

On the website: Copy pasting the same old prompts still work... but the 'flavor' of the text that pops out definitely feels worse.

I haven't seen any actual data showing a degredation.

AGI will very likely develop atop the mundane applications that AI will be utilized for at the onset - if it ever develops.

Regardless, mirage or not - as a mundane tool with realistic application this is still revolutionary

maybe they need some sort of profit center to keep themselves funded to reach their goals

Pessimism is more infective than optimism. It’s also easier to be a naysayer, since most new ventures do failed you can point and say you were right regardless of merit.

skeptical about AGI?

no, presuming that other people working on things are both cynical and deceptive; it's projection, imo.