dwroberts

Hungary had a revolution against Soviet control that was brutally suppressed and hundreds of thousands of people fled or were exiled. So yes, the Soviets were absolutely responsible for their communist dictatorship because they killed or removed anyone that did not comply

I did this (disabled AI features) and lost smart inbox and it honestly improved the experience: it made me realise how much rubbish I was subscribed to that I never read, so I unsubbed and blocked tons of stuff, and now my inbox is clear again

> Mostly capability failures that newer models no longer make: counting letters

Have LLMs fundamentally actually changed or improved to resolve problems like this, or have the big models been aggressively tuned / trained on specific examples so it’s less embarrassing?

Because I’ve noticed that Google’s stuff always seems to generate code to do counting tasks, even when not asked for, so I wondered if the LLM itself even generated the answer or it’s just being instructed to use a tool? (Which IMO is a workaround)

Lot of comments here taking this at face value, but isn’t it possible this was always a negotiation tactic on licensing costs?

Also wonder if this is eg MS offering a good deal at the eleventh hour to stop it going ahead, or if they called their bluff

Revenue by definition does not include operating costs

The models are also insanely inefficient and economically unviable to actually train and run. They’re doing the Uber style model of hoping they can keep afloat long enough to make other businesses go bust and then they can snap up their customers/market and amp up the prices

It is implied to be illegal despite Google claiming it followed all laws etc:

> Public reports cited by the LVV indicate that around 300 hectares of Finnish forest, roughly 420 football fields, had already been cleared before the project's environmental impact could be properly assessed.

Presumably with the environmental assessment being a prerequisite for a permit or something

> Was anyone under the impression that AI would, in its very first attempt at biological research, suddenly be able to get answers without having to need physical follow-through on its leads?

Why can’t it build a simulator for this stuff? It has textual and formulaic explanations of everything it’s working with. Seems like it should be less than a days work if it can solve all of math in a few hours?

Any computational bottleneck on running such a simulator should also be solvable right, Anthropic have already told us this is the singularity already

Isn’t that the point though - they’ve generated such strong results in mathematics, so why now are they delivering something so weak here?

Hype rather than realistic attempts to solve things is actually a valid explanation for that. They want to give the impression it’s a solve-everything machine even if it’s not well suited to areas outside of CS and math (or really anything where the logic of what is being done is not 1:1 with syntax, which is where LLMs are clearly excelling)

Fact check against what? Other facts it can find online that manipulate its behaviour into citing them?

Made a comment to similar effect - I lived near a bridge/underpass that was built at the end of the 1800s and it was pretty much an iron box. The assumption that these things need to be attractive or decorative just makes me think the writer hasn’t been around much in the UK

This is a really weird article if you’ve seen basically any non-decorative railway bridge anywhere in the UK.

Starting to make these WSJ stories of ‘UK is doing <whatever> wrong’ look a little forced?

Since it’s paywalled hard to tell if they included the actual facts about why it took so long too:

> The works in Botley Road were initially due to be completed in October 2024. Setbacks ranged from archaeological finds, the discovery of a World War Two grenade and essential work to replace part of the gas network

> any remaining unsoundness bugs are most likely in Lean's compiler or runtime (including the bignum library). It is very unlikely that the same bug also exists in the Rust compiler or runtime, so by checking a proof with both con-leche and con-ron, you gain a high level of protection against this class of bugs

An awful lot of references for arbitrary precision arithmetic are just repackaging Knuth’s explanations, so I wonder if there could in fact be a bug in common (Although you would like to assume every other piece of software using them would have also hit any issues by now, so maybe it is fine)

You don’t have to complete that puzzle to get a satisfying ending to the game

It reads like someone making up new terminology to make their work sound complicated. Seeing ‘organism’ and ‘superposition’ in a discussion of how NNs work reads strangely to me

This is the first of Blow’s games that I’m just not very interested in playing, and the fixation on its length and size is why. Being really long or having a huge completion time for the sake of it seems like a big step backwards compared to Braid and The Witness, and also philosophically literally the opposite of what he claimed made his games interesting. So I find this game and how it’s being marketed etc pretty odd

Two hosts they name in the post are Radist and Selectel

Why does it have the name nanolathe attached? What is that about? You can’t just steal an entire commercial game and give it a new name and pretend it’s your own thing

Also, a quickly vibecoded project doing something important to security, to be used by other applications, seems like a perfect way to drop a malicious backdoor (and maybe even provides the author plausible deniability when it’s found)

There is already a pre-print making the claim that the Navier-Stokes solution provided by OpenAI does not meet the criteria required by the prize (https://arxiv.org/html/2609.20803v1)

(and as noted on Wikipedia, the Clay institute still lists the problem as 'active' not solved https://www.claymath.org/millennium/navier-stokes-equation/)

It was a power plant full of fuel, why is this surprising?

> it cooked mathematics

Weird, these are all still here? https://en.wikipedia.org/wiki/List_of_unsolved_problems_in_m...

Given the context this made me assume extend is also some kind of a scam

‘Safe rust’ just being used to mean ‘no unsafe’ is kind of a shallow understanding of the language. You can write code without unsafe that is not really ideal at all eg abusing vector/slice indexing to create a kind of interior mutability that the borrow checker is blind to. Which as a C port I’m going to guess it probably ends up doing

The attributes on it seem irrelevant if it’s never called since it is presumably removed by dead code elimination?

Did you read what I wrote though? Sure it will immerse you - but the game also relies on visuals to send signals to the player about how their city is doing. If things are weathered and worn is that bad? Will I be able to tell the difference between ‘weathered’ and ‘abandoned’?

Also the visuals can be misunderstood to be communicating something - e.g. the dirty walls - does that mean my city is bad in some way? Is crime increasing or is the wall being stained by pollution?

You can't just make stuff 'look nice' without affecting how the player behaves

This has a strong AI psychosis feel. Absolutely no clue what I’m reading on any part of this site

Whether it takes milliseconds or seconds though, you are still doing it. You’re still just running a VM

> I see firecracker replacing docker on the horizon

I don’t think this is going to happen because they serve different purposes. Having to boot an entire OS inside a VM is a step backwards compared to containerisation. There are definitely use cases for isolating containers by running them inside a VM (see: Kata) but it generally ‘replacing’ docker, I don’t think that is on the horizon at all

(Also: you need KVM available, you need a rootfs to boot which will be larger than a container, it has no built in support for mounts or really any kind of communication with the host unless you explicitly set up networking etc for it)