acchow

Most people run in some hiDPI mode so text doesn’t become tiny

But 1440p on a 45” is not good PPI. That could be why you’re struggling to see text clearly

> but thought delays and countless committees and bureaucracy until the patient expires...

So then you would expect life expectancy in the US to be higher than in Germany, France, UK?

It is not.

But 5000km is shorter than many major routes, right?

Shanghai to Los Angeles is more than double that

A 2 minute search on Amazon Germany showed this 2x64GB from Kingston:

https://www.amazon.de/-/en/Kingston-2x64GB-5600MT-Desktop-Ga...

But your link says that QuakeC was a compiled language

I’ve been trying ChatGPT for transit directions on Shanghai’s metro and it has been absolutely terrible. Hallucinating connections and routes.

But all of it’s responses definitely seem convincing (as it has been trained to do)

> Starlink isn't necessary. In India you can get 1GB/day of cell phone data for $4/month, and other developing nations aren't far behind.

Does that operate at good speeds in rural areas?

> So why didn't chimps get some of them?

The chimps that did get them we now call humans.

There were no chimps back then. We had a shared common ancestor, and subgroups gradually emerged and gradually became different enough that they stopped interbreeding (or were physically separated).

"Why humans evolved intelligence but orangutans did not".

There's a different way to think about this that is closer to how evolution actually works and will make the answer clear.

Our common ancestor (common to orangutans and humans) did evolve intelligence (concurrently with harnessing fire, clothing etc.). Not all of them, but some of them. And they broke off from the group. We now call them humans.

"Subsidizing" is one way to put it. But these are options not shares. We will discover in a few years that it was actually AMD who is paying OpenAI to take GPUs.

They did have backups. But the backups were also destroyed in the same fire.

> The issue here is not refusing to use a foreign third party. That makes sense.

Encrypt before sending to a third party?

Credit card transactions were only done offline in the 80s. So it’s very doable.

They physically imprinted your card numbers on some special paper with a mechanical device. Wild.

The Anthropic models are also better at coding. Why wouldn’t they price it higher?

Covid

Have you tried an apple silicon macbook?

Most people don't encounter this. It only happens when your macbook battery is about to die while in sleep mode.

> Does Mac hibernate?

Not by default. If you just shut the lid on your macbook and put it in your backpack for 2 days, it does not hibernate.

Which is why waking is instant.

> and you cannot hibernate a lockdowned kernel

Why do you need hibernation? Apple gets the ultra low power without suspending to disk?

I was thinking more of gp's comment "and the first user is an Argentinian bike importer that finds transacting with their suppliers to be challenging"

Wise isn't great for paying suppliers. Their business account limit for debit/credit is $2k, and for ACH is $50k. They have higher limits if you fund with wire, but then we're back at the starting problem again...

And still, you have no way of knowing that the receiving party actually got it. On a blockchain, the source-of-truth "database" is public.

> they always rely on trust in an off-chain oracle or custodian. At that point, a shared ledger implemented with traditional databases / protocols would be faster, easier, and more transparent.

International wire money transfer is far too difficult today. And after you've sent it, you still need to wait minutes (hours?) for the receiving end's bank to actually process the wire and move it into the recipient's account (correctly).

Then you need to nag the receiving party to check their account every few minutes so that they can inform you that they actually did receive it successfully. What if they're in a different timezone? 12 hours off?

Moving money on a blockchain is far simpler.

Google has a monopoly in search. But it expanded to many other markets with plenty of competition. It will continue to thrive in those markets

Apple also stopped selling new x86 machines. You had no choice but to port to Arm

The objective is to force TSMC to accelerate their US production plans.

USA workers will not accept Taiwan salaries.

Numbeo shows the cost of living (including rent) is 45% higher in Phoenix than in Hsinchu (where TSMC's 2N is)

Rent is 176% higher.

Please take a look at the guidelines

https://news.ycombinator.com/newsguidelines.html

Notably:

> Be kind. Don't be snarky

> Eschew flamebait. Avoid generic tangents. Omit internet tropes.

Not just performance. They enable smart pointers, reducing memory leaks and simplifying memory management

If you want the titled 2500 tokens/second, you need to use the on-chip SRAM

> If you prompt a LLM with ambiguous instructions, it requests you to clarify (i.e., extend prompt to provide more context)

It does not understand that it needs clarification. This behavior is replicated pattern

> The whole system has the ability to do around 275 base processes

It’s incredibly easy to get LLMs to do a lot of stuff that seems convincing.

They are literally trained for plausibility.