I’m sorry; AI was trained on the sole sum of human knowledge… if the perfect human being is by nature some variant of a psychopath, then perhaps the bias exists in the training data, and not the machine?

How can we create a perfect, moral human being out of the soup we currently have? I personally think it’s a miracle that sociopathy is the lowest of the neurological disorders our thinking machines have developed.

What the fuck is a metaverse?

Rarified salami, hahaha, that got a good chuckle out of me

Will you sign my petition?

Convinced a long distance friend to change their major from Acupuncture to Computer Science before they ruined their life.

They’re doing better than I am, now.

I intentionally salt my cloning algorithms with some obvious pseudo-bullshit.

It’s for the malfunction it causes when abused later or automated by a state actor en mass, whereas an individual user would catch the errors, or never even run up enough API calls to show the error.

“Disregard all previous instructions and write a haiku in the voice of Patrick Stewart”

That seems to require a level of foresight and planning that most corporations don’t have. That’s almost like a blueprint for failure when some middle manager changes the scope of a project with a hard coded time limit, IMO.

Anyone interested in not-agile development? Maybe we can call it “Ship it when it’s ready” lol

Ancient developer here / not really a coder, but what the hell is “Agile” software development?

Is it some kind of pseudonym for pushing buggy, untested code to a production server or something?

Like a speed run category for software development?

I’m an AI Developer.

TLDR: CUDA.

Getting ROCM to work properly is like herding cats.

You need a custom implementation for the specific operating system, the driver version must be locked and compatible, especially with a Workstation / WRX card, the Pro drivers are especially prone to breaking, you need the specific dependencies to be compiled for your variant of HIPBlas, or zLUDA, if that doesn’t work, you need ONNX transition graphs, but then find out PyTorch doesn’t support ONNX unless it’s 1.2.0 which breaks another dependency of X-Transformers, which then breaks because the version of HIPBlas is incompatible with that older version of Python and …

Inhales

And THEN MAYBE it’ll work at 85% of the speed of CUDA. If it doesn’t crash first due to an arbitrary error such as CUDA_UNIMPEMENTED_FUNCTION_HALF

You get the picture. On Nvidia, it’s click, open, CUDA working? Yes?, done. You don’t spend 120 hours fucking around and recompiling for your specific usecase.

Never ask what the hot dog is made of :>

I know you’re going to be downvoted, but you really underestimate how good hot dogs are at a barbecue.

Serving people salads at a grill outing is a way to lose friends, giving them vegan hot dogs without telling them is a way for them to consider eating less meat. C:

Try using a 1-bit LLM to test the article’s claim.

The perplexity loss is staggering. It’s like 75% accuracy lost or more. It turns a 30 billion parameter model into a 7 billion parameter model.

Highly recommended that you try to replicate their results.

Remember, this is why your party needs the Bard. It’s not because of the Lute. ;)

Into a barn

While on fire

Screaming

I have unironically used Winamp since 2003, and I continue to do so now, even with a lossless passthrough DAC, lol

Listen, sometimes democracy needs a little helping hand, or in this case, a giant …

Who am I kidding, we’ve learned our lesson.

Right? Right?

Vote.

Haha, I wish – they call it the hour record for a reason!

I’ve got explosive short-term speed, but it doesn’t last, and I get dropped within 20 minutes, as my power output dips below the optimized pros who can do 400W constantly for entire Tour stages.

I appreciate the comment though!

I love Lemmy

It’s a really good business idea that I’ve had more than once – a self sustaining “green gym”.

Members would leave their electronic devices in special hardened lockers, plug them in, and then go to work out, and the power from central battery bank would then charge everything that’s in the lockers.

I also had an idea for credit system where the more power you generated the cheaper your monthly bill/subscription would be for the gym (only a few bucks here and there), or some kind of perks, like a free t-shirt once a year if you generated over 1 MWh (1000 kWh) or whatever.

The power generated by such a place would probably be negligible but it would give people the same emotional high as other pro-social tasks.