https://gdpr-info.eu/art-22-gdpr/#:~:text=The%20data%20subje...

What it does incentivise is pushing more usage in case there’s a reset-generating additional demand. Probably sounds good in the boardroom that token consumption is up X% this quarter.

Unlikely to last forever, but for now the goal is to beat Anthropic. Up until the Mythos/Fable debacle, I’d have said Anthropic were leading, but if you’ll forgive the pun, I think they flew to close to the sun, and when Sol came out, they’d shot their “Mythos/Fable” shot and it turned out to be… good, but not revolutionary.

If you just want to use ATProto for your own small app, there are ways to do it on more constrained hardware. The more usage it gets, the more it scales up.

> On ActivityPub and Nostr you develop all kind of applications already, it is not only Twitter clones, but also Instagram clones, Reddit clones, on Nostr there is even the resurrected Vine (https://about.divine.video/).

Same! You can explore the (unfortunately named, given the point I'm trying to make :)) 'Bluesky directory' to see apps which use ATProto: https://blueskydirectory.com

Note that most of these have nothing to do with Bluesky.

> And it is not trivial to switch from the Bluesky did:plc (which you cannot run in write-mode on your own and need to use the official Bluesky owned) to did:web, which you can run on your own

That's true, but it's not trivial to run your own Mastodon instance either...

What do you mean? Individual apps can choose to censor I guess, but you can fairly trivially run an AppView with no filtering whatsoever if that's what you fancy.

> - sure, you have your data in your PDS, but it is de-facto useless if the central identity server of Bluesky goes down

That's not correct. You can use plc, which Bluesky operates the largest directory for, but you can also use web, which completely bypasses Bluesky all together.

> - ActivityPub and Nostr solve this already without being owned by a company

What's 'this' here?

> - you cannot easily and cheaply have an ATProto instance. AP and Nostr solve this already

The terminology here of 'ATProto instance' doesn't really make sense. If you're talking about a PDS, you can use one of thousands of free PDS hosts, or host your own for free on Cloudflare (https://cirrus.earth/) or on your own hardware (there are PDS implementations which fit comfortably onto a Raspberry Pi, or an old phone, etc.)

Out of ten items on your feed, five would be adverts, three are reposts of some crappy influencer’s video or a meme or something, one is a friend’s marketplace listing, and if you’re lucky, the last one is a photo actually taken by a friend who you actually know.

https://alphatab.net/docs/alphatex/introduction

Not quite MD, but fairly easy to learn.

So once you’ve solved this problem, you’ll need to ensure there’s no software running which can capture the screen and/or keyboard inputs.

Then make sure that every peripheral is running firmware which you know isn’t compromised.

Then make sure the hardware itself is not compromised.

Then make sure the user’s screen is not being recorded. Or visible on any CCTV camera.

There’s a line between utility and perfection. Everything in the real world lives somewhere to the left of “perfection”.

The article is not wrong that there’s a threat model that makes our current ways of doing E2EE less than perfect, but it doesn’t impugn the actual E2EE implementation at all, and is not much different to saying “your data is still not secure because someone could take screenshots!”

What AI unlocks is contributions from folks who are not at all involved in the project, and creating a PR is no longer enough to clear the gate of “this person is at least somewhat interested in the success of the project”.

AI is a force multiplier when wielded properly, but as an OSS maintainer, it’s easy to drown in prolific, low value “contributions” from folks who don’t care about the project.

Not sure the “causal connection” is fully established.

The real question is where’s the cut-off point between accuracy and utility.

Remember: a second human opinion can also be wrong, and even a wrong opinion can still be useful (especially in medicine where differential diagnoses are a common practice - if the LLM gives you a useless opinion, you rule it out and move on).

I don’t think it’s particularly unreasonable to think that an LLM would have enough literature, or enough reasoning ability, to be able to generate a plausible interpretation of the data. A human can then review and say either “yeah that’s clearly not the case here” or “hmm, actually that could explain it, maybe we should order another test”.

On top of that, local retailers have to comply with EU regulations etc. while Chinese imports are notorious for not following safety/recycling standards etc., which unbalances the playing field away from local retailers anyway.

The more training data, the more questions it can answer with a reasonable degree of probability of accuracy.

Throwing away a potentially useful analysis just because it’s probabilistic seems a bit like throwing the baby out with the bath water.