drcode

Let's hope you are right, and that the limitations of AI will never improve, and that all of us get to live out our full natural lifespans without AI xrisk

Wealth isn't about "the best software tools"

Wealth is the time required to make great art- Every great artist needs this.

Currently, some artists are able to make a living from their art, and can spend the time and effort to create great art without being independently wealthy- but that is going to become increasingly difficult

because I predicted that AI will get better in future months?

As far as I understand, coding ability of AIs is now driven mostly entirely by RL, as well synthetic data generated by inference time compute combined with code execution tool use.

Coding is arguably the single thing least affected by a shortage of training data.

We're still in the very early steps of this new cycle of AI coding advancements.

Yes, independently wealthy musicians and authors will still be able afford investing the time required to make great art

> {In late July of 2025} all LLMs hit a ceiling of complexity beyond which they cease to understand the code base

Fixed that for you

sort of, except I think the future of llms will be to to have the llm try 5 separate attempts to create a fix in parallel, since llm time is cheaper than human time... and once you introduce this aspect into the workflow, you'll want to spin up multiple containers, and the benefits of the terminal aren't as strong anymore.

I was kinda pissed when my local mall got a "barista robot", and it asks for a 20% tip when you swipe your card

Let me introduce you to Hard 2632, a device for 32 byte demos: https://xayax.net/hard2632/

damn, that's a crazy process- thanks for the video link

ok that's a fair point

That's seems silly, it's not poisonous to talk about next token prediction if 90% of the training compute is still spent on training via next token prediction (as far as I am aware)

thanks, I'll investigate my local gym!

update: they use trueblue

PSA: Most modern gyms have "autobelay" devices that let you climb on your own without a partner. This makes gym climbing a super fun and accessible exercise anyone, even beginners, can do by just showing up to a gym at your convenience.

(If you're a beginner you should still take the 1 hour class first and you will have to pass a belay test. And yes, if you can make the schedule work out with a friend so can belay each other, that's even more fun)

the madness of cutting hundreds of gears by hand, many with a 45 degree bevel, I couldn't even imagine.

madness!

> How is AI going to make its own chips and energy?

Pay naive humans take care of those things while it has to, then disassemble the atoms in their human bodies into raw materials for robots/datacenters once that is no longer necessary

I suppose there is an equilibrium, where sites that penalize these types of crawlers will also get less traffic from people reading ai citations, so for many sites the upsides of allowing it will be greater than the downsides.

It's a crazy accomplishment really, unimaginable how safe US commercial airplanes have been in the last two decades

So sad that streak finally ended

longevity for the AIs

eh, they will add a quantum-resistant signature scheme (already a well-understood thing) then people can transfer their funds to the new addresses before it is viable to crack the existing addresses

To clarify, this is specific to "alphabetic" writing, cuneiform/hieroglyphs are older.

Another "LLM skeptic" arguing it is impossible for LLMs to ever reliably answer a certain type of prompt, because they can't answer it reliably in 2024.

Every previous skeptic making this argument has been proven false so far

Per page 234 of FARS, drivers are classified as occupants https://www.ire.org/wp-content/uploads/2018/12/USERGUIDE-201...

Companies are hiring fewer people that understand a technology that enable companies to hire fewer people- News at 11.

Zig believes that all code needs to be explicit, to prevent surprises- You never want code that "just executes on its own" in places you may not expect it. Therefore, if you want default arguments, you have to perform some action to indicate this.

my experience is that o1 is still slightly better for coding, sonnet new is better for analyzing data, and most other tasks besides coding

> in general it's completely normal to assume someone wearing earbuds is listening to you

Just because you emphatically state that something is normal, that does not make it normal

Agreed, it seems like NVIDIA would be happy to make whole-wafer chips if it seemed like a good play.

My guess is there are a lot of bespoke limitations that the software has to work around to run on a "whole wafer" chip, and even companies that have 99% similar designs to Nvidia already are struggling to deal with software incompatibilities, even with such a tiny difference.

turning the earth into paperclips is not gonna happen immediately, so we can safely ignore that risk

we'll probably create artificial superintelligence in the next few years

when that happens, it likely will not go well for humans

the specific harm is "human extinction"