IsTom

It sounds like it was a bug caused by pushing on performance/safety ratio, it could easily have been avoided by adding enough mutexes in the code, but for performance reasons (I assume) they don't do that. To write it like that in Rust you'd have to use unsafe blocks and then you don't gain much from the typesystem in this case.

If you believe that I've got a halting problem oracle to sell to you.

I might have learned all this too long ago and lost touch with how it's like to learn this, but: how is learning about using pointers in C different from using object references and array indices in python/js/java?

The point is that you need multiple plantations like this to make enough feed for the cows to make equivalent calories/proteins.

TBH fortified yeast flakes are a pretty good spice.

Why couldn't they, if they've got DNA and ribosomes?

> but also gave us Amazon, Google and 2020s Microsoft.

The major fault I see here is that people didn't forsee that the default GPL should have had AGPL's clauses.

> if the productivity boost is high they will use it anyway if their rules keep crippling their own models

Individuals maybe, companies won't and that's where most of money is at.

Back then OOP was all the hype and inheritance was cool.

How is 30B smaller than 27B?

Adding too many tests was a real issue before LLMs. You end up in situations where you when you add a feature there's 200 (out of e.g. 10k) tests that fail and you have to figure out which of them should fail and you need to fix them and which are actual bugs.

With the "It's not X, it's Y" at the end I sure hope that it's the case.

And not to be confused with its plural form "opera".

> So many audio-to-digital conversions used cheaper anti-aliasing filters, with a narrower pass-band, which could be significantly lower than the 20 kHz promised by the CD technology. The CDs mastered in this way had a worse quality than claimed.

That wouldn't result in "tinny" sound though as the article claims.

> Record companies accordingly used sampling rates just dense enough to keep most listeners from complaining, and called it good. It was good for their profits, sure, but it made the music sound flat and tinny.

Were there CDs with lower sampling rates than the standard 44.1kHz? I could understand complaining about low quality DACs or something, but if 44kHz is not enough for you then you're either a child or a bat.

Until they apply what that did to Github's uptime

Internet walled gardens are all opt-in and still they've made the web a worse place.

Still there's not a single example.

I think it's enough that a single layer fits on each chip if you can daisy-chain them with good interconnects.

Hasn't this been happening for a while now?

Are you saying that Hurd will also kill Google/Linux?

It might also be that the security sandbox was haphazardly vibeslopped together.

There's plenty of tools that can cause harm without having to be creative to use them.

I think it's just regular corporate speech. LLMs do this, because they've learned on this kind of posts.

We'll yet see where the wind blows, but there's a real possibility of copyright issues for machine-generated code down the line.

If I wanted to be uncharitable, the question of "how much slower you'd be without AI" compares against people's AI-less productivity from 2026, which could have become much less than 2025 AI-less productivity, if hypothetically AI use makes people helpless and deskilled.

Ah, that makes sense, thanks.

But won't your boat displace more water by lowering a bit if you've got additional mass on board?

If you've got X GB of weights in slow access memory (be it RAM vs VRAM or SSD vs RAM) and Y GB of fast memory then no matter what, if you want to use them you'll need to transfer X-Y GB and will be bound by memory throughput. You can try to reduce number of activated weights, but how much can be gained that way is speculative so far.

I think it speaks about the current state of github that I had to scroll 12 projects to find the first entry not related to AI.