dist-epoch

That's a skewed perspective.

Discord always deeply cared about their users. Which is why nobody is able to get even close to replacing them.

They are adding user requested features and improvements all the time. They might not be what you consider "improvement", but maybe you are not really their target market.

Until someone sits in front of them

https://www.youtube.com/watch?v=Srytx_01MW8

Maybe it's something like some of the missiles have no warhead on them, and you don't know which ones.

I don't think you understand what the word proof means

LLMs can find endless combinations of edge cases that are a real bug. Addressing them requires thousands more lines of code. Which creates another set of new edge cases to combine.

At some point you realize recovery is better then prevention, let the bugs happen, but recover from them gracefully.

> LLM may be a great technology, but it's probably not a good model for how humans learn and use language.

but from all models, it's the closest, "all models are wrong, but some are useful"

Only those who never traded the stock market for real (like financial journalists) believe that you can attribute news to a market event.

"The market went down because the Fed increased the rate"

So why does this post have 200 upvotes then?

Nobody said the prompt is 5 lines.

It could be a 2000 line prompt (a spec).

Prompt injection is very much not real anymore.

This site doesn't do jokes

As the motto says, "you can just do things"

Such a rewrite would have been unthinkable 2 years ago, but now it's table stakes.

Modern inference stacks use different GPUs for decode versus prefill. This is a GPU suitable for prefill.

Apple M6 has 80-100 TOPS.

Qualcomm Cloud AI 100 Ultra has 800 TOPS, almost 10 times more.

Refreshing, compared to "our journey" and "our mission" slops

An attacker might replace a package with a previous one which has a vulnerability. Downgrade attack. Unclear if the apt client will check if version is the expected one after requesting a package.

> NVIDIA says that the new DGX Spark 64GB is starting at $4,999

> NVIDIA has now confirmed an official starting price of $2,399 for RTX Spark laptops

https://shattered.io/nvidia-rtx-spark-mediatek-oems-price-20...

https://www.servethehome.com/nvidia-dgx-spark-64gb-launched-...

if any of your SSH key files are bigger than 500 bytes, you should rotate them

I'll do it for free. I get a thrill from rebooting servers and hoping everything still works.

Because only a finite amount of time passed since the Big Bang, and all speeds are finite, so it could have grown only a finite amount.

It's not weird at all, they know people will buy these laptops, put linux on them and stash them in data-centers, so they are trying to prevent this.

You know what would be more efficient?

Shipping your drives to Sam Altman. If everybody did that, he could then deduplicate the pictures of the same thing that multiple people took.

Applied calculus doesn't need infinities. No computer can represent a real real number. Only truncations of them, only finite versions of them.

> Many other extremely beneficial devices for extremely high value use cases

If the other use cases are extremely beneficial and extremely high value, surely they can happily pay $1000 for 64 GB of RAM?

And yet the vast majority of physicists treat infinity in the equations as a sign of them being incomplete, and dismiss the possibility of the universe being infinite in any way

I would like to point out that there is also no proof of God not existing either

indeed, the purpose of theoretical computer science is as a filtering and ranking mechanism for awarding tenure in universities

Sensible position in a finite universe

And yet you ask it 3 times to not do a thing, and it does it in a 4th way. A human doesn't fail in this way. And I'm not talking about memory here, single context.

So yeah, it can do everything, but fails in trivial ways around the doing part.

Something like mirror neurons.

Some structures in Astra get activated by the statistics of the Qwen weights, and these structures are associated with "breakout"

there are tiny chip size wireless receivers for small distances like this, you would need some sort of window through the metal, or a tiny antenna inserted through the ball shell and the wireless covers just the inside

or just ultra-thin wires connecting the memory to the outside. the ultra-thin wire will just snap and it's too light to cause damage by whipping around

the big problem is kinetic impact, fire/fuel/water would be easy to solve, for example fire requires oxygen which will have difficulties getting inside through a tiny hole, and I'm sure there is some metal alloy which doesn't melt in typical plane fuel fire temperatures