gnatolf

While I don't disagree, memory certainly was more of a restrictions on us humans than it is on llms. Therefore, the answer may not be as obvious as it seems. We build abstractions to reduce (memory) footprint of features, right?

All over Germany, and it's been around much much longer than the fear of having something slipped in your drink.

Which, to be fair, overall probably still saved quite some time. The compile times alone would've meant they wasted so much more time by repeated earlier checks.

Whether a looming deadline changed the perception about that, we don't know ;-P

Interestingly enough, precise search is on the way out.

'liquid metal' sounds cool. It's probably a metallic glass. I super dislike that it seemingly will be synonymous with the brand name by Apple even though that stuff has been around for decades.

Not that there are particularly many places where this is used - mostly because it really is just very expensive. In the awesome position that Apple is in, economic feasibility is so much easier to achieve, with like tens of millions of guaranteed parts to be preduced.

Well, there is a financial 'sink' - stockpiles and ammunition or other non-reusable military gear are basically the definition of money 'destroyed'. Their political value is almost non-existent actual money. If any, at all.

Sorry but isn't the bottleneck then simply to do even relevant things? Like how much of a qualified backlog do you have that your pipeline does not run dry?

So let's put things we're interested in in the benchmarks.

I'm not against pelicans!

Good point. So much functionality gets commoditized, we have to move goalposts more or less constantly.

While this is funny, the actual race already started in how companies can nudge LLM results towards their products. We can't be saved from enshittification, I fear.

More specifically regarding spec-driven development:

There's a good reason that most successful examples of those tools like openspec are to-do apps etc. As soon as the project grows to 'relevant' size of complexity, maintaining specs is just as hard as whatever other methodology offers. Also from my brief attempts - similar to human based coding, we actually do quite well with incomplete specs. So do agents, but they'll shrug at all the implicit things much more than humans do. So you'll see more flip-flopped things you did not specify, and if you nail everything down hard, the specs get unwieldy - large and overly detailed.

Everybody feels like this, and I think nobody stays ahead of the curve for a prolonged time. There's just too many wrinkles.

But also, you don't have to upgrade every iteration. I think it's absolutely worthwhile to step off the hamster wheel every now and then, just work with you head down for a while and come back after a few weeks. One notices that even though the world didn't stop spinning, you didn't get the whiplash of every rotation.

That is a wobbly assertion. You certainly would need to run the same compiler, forgo any recent optimisations, architecture updates and the likes if your code has numerical sensitive parts.

You certainly can get identical results, but it's equally certainly not going to be that simple a path frequently.

From smaller to larger nitpick, there's basically something wrong with all of the first 15 or so of these drawings. Thanks for agreeing :)

Absolutely. A technically correct bike is very hard to draw in SVG without going overboard in details

Yes, even if you create a single person account, you create an 'organization' to be billed. That's the whole confusion here. Y'all seemingly don't have an account at anthropic?

Craftsmen will have a resurgence, that's probably a 'leveling up' in terms of resilience against AI takeover. There's just no way of automating quite a few of the physically effective crafts.

Love it when it circles around a minor issue that I clearly described as temporary hack instead of recognizing the tremendously large gaping hole in my implementation right next to it.

I was somewhat confused and irritated by the lack of a clear frontrunner crate for XML support in rust. I get that xml isn't sexy, but still.

For me, the shear length of uuids is annoying in payloads of tokens etc. I wish there was a common way to abbreviate those, similar to the git way.

If only one could construct a macro to solve the boilerplate of AsRef<str> etc ;)

Which confusingly are linked to on the 'prev' link at the bottom of each part...

Specifically not in the current case where almost all of that GDP growth lines the pockets of the already-wealthy.

You're assuming that given the collection of simonw's publicly available blog posts, the creativity of those combinations can't be narrowed down. Simply reverse engineer his brain this way and you'll get your Xs and Ys ;)

Current examples would be drag racing cars that have motors that are designed and used in a way that they only survive for about 800 total revolutions.

Yea, I know. Their volume may look impressive, the actual amount of material is quite small and 'burying' that absolutely non-toxic stuff isn't any problem.

'recyclable' is such a vague term. E.g. radiation-affected typically easily recycled materials are very hard to deal with (think e.g. pipe steel from power plants) and are effectively non-recyclable, instead of close to 100% recyclable, as their non-contaminated counterparts.

Opposed to that, battery recycling is mostly hard to deal with in terms of economics, and admittedly the chemistry involved is complex, but at least from a technical point of view, plenty of solutions are available - and the tech is coming in relatively quickly now that the demand is there (remember, first generation EVs are just now getting closer to EOL).

It's slightly amusing that recycling of a wind turbine is treated as if it was a big deal - yes the laminated rotor parts can't be part of circular economies, but the total material amount of this laughably small. All the metal components are very easily recycled.

Curious what's the argument for/against that here. I agree it's at least unusual.

Which inexplicably isn't know for soft landings

I am not as optimistic that locally run models will be able to compete anytime soon. And even if, the price to run them means you have to buy the compute/gear for a price that is likely equivalent to a lot of 'remote' tokens