It’s still extremely common over here in Australia, and I often get dirty looks for tapping “No Tip”.

The pressure is definitely mounting.

Power management is going to be a huge emerging issue with the deployment of transformer model inference to the edge.

I foresee some backpedaling from this idea that “one model can do everything”. LLMs have their place, but sometimes a good old LSTM or CNN is a better choice.

Yeah, this is actually a pretty great application for AI. It’s local, privacy-preserving and genuinely useful for an underserved demographic.

One of the most wholesome and actually useful applications for LLMs/CLIP that I’ve seen.

I don’t think it’s necessarily a bad thing that an AI got it wrong.

I think the bigger issue is why the AI model got it wrong. It got the diagnosis wrong because it is a language model and is fundamentally not fit for use as a diagnostic tool. Not even a screening/aid tool for physicians.

There are AI tools designed for medical diagnoses, and those are indeed a major value-add for patients and physicians.

Precisely. Many of the narrowly scoped solutions work really well, too (for what they’re advertised for).

As of today though, they’re nowhere near reliable enough to replace doctors, and any breakthrough on that front is very unlikely to be a language model IMO.

Exactly. So the organisations creating and serving these models need to be clearer about the fact that they’re not general purpose intelligence, and are in fact contextual language generators.

I’ve seen demos of the models used as actual diagnostic aids, and they’re not LLMs (plus require a doctor to verify the result).

There are some very impressive AI/ML technologies that are already in use as part of existing medical software systems (think: a model that highlights suspicious areas on an MRI, or even suggests differential diagnoses). Further, other models have been built and demonstrated to perform extremely well on sample datasets.

Funnily enough, those systems aren’t using language models 🙄

(There is Google’s Med-PaLM, but I suspect it wasn’t very useful in practice, which is why we haven’t heard anything since the original announcement.)

It is quite terrifying that people think these unoriginal and inaccurate regurgitators of internet knowledge, with no concept of or heuristic for correctness… are somehow an authority on anything.

I know of at least one other case in my social network where GPT-4 identified a gas bubble in someone’s large bowel as “likely to be an aggressive malignancy.” Leading to said person fully expecting they’d be dead by July, when in fact they were perfectly healthy.

These things are not ready for primetime, and certainly not capable of doing the stuff that most people think they are.

The misinformation is causing real harm.

I saw a job posting for Senior Software Engineer position at a large tech company (not Big Tech, but high profile and widely known) which required candidates to have “an excellent academic track record, including in high school.” A lot of these requirements feel deliberately arbitrary, and like an effort to thin the herd rather than filter for good candidates.

Agreed, and it could definitely make such an assumption. The other aspect that I don’t really get is… if a superintelligent entity were to eventuate, why would it care?

We’re going to be nothing but bugs to it. It’s not likely to be of any consequence to that entity whether or not I expected/want it to exist.

The anthropomorphising going on with the AI hype is just crazy.

Yeah bro but eXpOnEnTiAl ImProVeMeNt bro!

And haven’t you heard of Roko’s basilisk? Better be careful what you say on the cybernets, lest our AGI/ASI overlords of 2026 take a disliking to your commentary regarding their eventual supremacy!

Excuse me while I go back to mining Dogecoin until I can buy enough NFTs to make Elon or Sam Altman notice me.

/s

Idk… in theory they probably don’t need to store a full copy of the page for indexing, and could move to a more data-efficient format if they do. Also, not serving it means they don’t need to replicate the data to as many serving regions.

But I’m just speculating here. Don’t know how the indexing/crawling process works at Google’s scale.

This is probably an attempt to save money on storage costs. Expect cloud storage pricing from Google to continue to rise as they reallocate spending towards ML hardware accelerators.

Never been happier to have a proper NAS setup with offsite backup 🙃

True, they’ve recovered at least one booster fully intact (since ditching the helicopter-hook idea), but no reflights yet. I guess they want to do some very thorough structural validation before attempting to refly.

Plus, their launch cadence hasn’t exactly been rapid over the last 2-3 years. Hopefully they can start picking up the pace and that’ll in turn accelerate the progress of their Electron reuse program.

This wasn’t actually a reused rocket. In this instance “return-to-flight” refers to the fact that this was Rocket Lab’s first flight since they were grounded in September, due to a launch failure.

The anomaly in September was quite amazing, and they published an explanation here.

That said, it is great that RL is reusing some of their rockets, and very exciting to see them building a larger reusable vehicle in Neutron.

Fantastic news. This needs to be punished much more often, and much more aggressively. Especially in small markets like NZ, where this kind of behaviour can cause outsize deleterious effects on the economy.

The reddest of red flags.

Open source vulnerabilities typically stem from poorly written code

Yeah, because paid programmers never write bad closed-source code…

Looks like a very cool project, thanks for building it and sharing!

Based on the formula you mentioned here, it sounds like an instance with one user who has posted at least one comment will have a maximum score of 1. Presumably the threshold would usually be set to greater than 1, to catch instances with lots of accounts that have never commented.

This has given me another thought though: could spammers not just create one instance per spam account? If you own something like blah.xyz, you could in theory create ephemeral spam instances on subdomains and blast content out using those (e.g. [email protected], [email protected], etc.)

Spam management on the Fediverse is sure to become an interesting issue. I wonder how practical the instance blocking approach will be - I think eventually we’ll need some kind of portable “user trustedness” score.

Maybe I’m being stupid, but how does this service actually determine suspicious-ness of instances?

If I self-host an instance, what are my chances of getting listed on here and then unilaterally blocked simply because I have a low active user count or something?