d1sxeyes

I think Sol is pretty much there already

It’s a false idea that a reset “gives you more”. You don’t get any more tokens post reset than before. If you were consuming faster than your allowance before the reset then you probably end up with some free usage, but most people are not likely so far over their pro-rata limit that it’s a huge cost for OpenAI.

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.

Permafable got announced earlier this week, but definitely the slot machine feel is real.

Effectively what I meant by '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', but appreciate the top-up!

The idea of an 'instance' is not really clear. If you want to run a social network on ATProto, you do need the relay and app view, and those are indeed heavy requirements, but the architecture is not the same, and it depends what you're actually trying to achieve.

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...

> - content not in line with the political left is censored

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.)

You can’t really have an asynchronous phone call though…

It’s gone, but it was already too late.

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.

AlphaTex?

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

Not quite MD, but fairly easy to learn.

> A cryptosystem is incoherent if its implementation is distributed by the same entity which it purports to secure against.

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!”

The key problem is that traditionally, OSS contributions were self-selecting. Basically, to create a PR, you had to be invested in the project. To create a contribution of value, you had to understand the codebase, the conventions, engage a little with the project, and generally the folks doing it are doing so because they like the project, or because they are scratching a specific itch they have etc.

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.

(2014)

Sorry for challenging here, I don’t remember the actual study being linked, and it seems I’m now paywalled and can’t check myself: did he do the same experiment with folks wearing sunscreen? Did he control for relaxation effects? I’d expect both blood pressure to go down and nitric oxide to go up as folks relax, e.g. in a sunny garden or on a tanning bed.

Not sure the “causal connection” is fully established.

Yes that’s my point. An LLM can clearly accurately predict obvious cases, so it’s reasonable to assume that it can predict less obvious cases with somewhat less accuracy.

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”.

For phone cases, maybe, but not for everything under 150 EUR…

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.

Not quite. An LLM generates text that would likely follow. The sky is… “blue”. A patient in pain with a bone protruding from their shin has a… “broken leg”.

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.

It’s not a “middleman rent seeking”, it’s protectionism. If lower cost production is available elsewhere in the world, there are three options.

1. Stop producing locally. Allow the market to take care of it.

2. Deregulate minimum wages to allow local businesses to price locally produced goods competitively.

3. Impose a tariff on incoming goods to protect local producers.

Which is your preference?

Surprised you don’t think HB is extremely popular. Perhaps you’re outside of Europe?

Thomann own an extremely popular brand, Harley Benton, which (obviously) sells S-types.

I think it’s more likely that this is about the ST series than about the Silver Sky.

The first two reported timestamps are 16:00 and 16:30 which both fall conveniently on the half hour. This leads me to conclude that these are guesses, and could be either early or late. It seems most likely that the mother was not exactly aware of the time the child left the house. There is an incentive to minimise (to avoid it looking like a failure of supervision).

What we are certain of is the 16:44 call to the EMS, so you're right, 134 minutes is the lower bound. However, it's not unreasonable to assume that the parents did not immediately call on noticing the child missing, so the gap between 'noticing missing' and 'calling EMS' is real and non-zero.

The gap between 'leaving the house' and 'noticing missing' is something I'm less clear on: how did the parents 'notice' the child was missing when he was knowingly allowed to leave the house earlier? There's still a non-zero gap between 'last sighting' and 'realisation that he's missing' (i.e. he definitely wasn't missing at ~16:00, when the mother saw him leaving the house... what changed so that the parents believed he was missing at ~16:30? Was he supposed to check in with Dad and missed the check-in, etc?)

Either way, a difference of ±10% doesn't really make that much impact in interpreting the results.

There are two key problems with social media. The first is a people problem. Predators, bullying, etc. This is a difficult problem. The furthest I got in ideating on this was some kind of graph-growth limitation: some proof of physical co-presence or chain-of-trust (i.e. approved by a trusted adult) required for friending.

The second is a product problem. Advertising, the 'algorithm', etc. For that, I do have a suggestion: force all social media sites to have ad-free versions which apply to the accounts of under eighteens, and attempt to reduce 'virality': no like buttons, ways to discourage reposts of original content, etc. so the timelines become less engagement driven.

It's also illegal in the EU to use data for purposes other than that which it is collected for. A company can't retrospectively use network logs and badge logs to track office attendance unless they specifically collect it for this use.

It would be very challenging for companies to retroactively change this without it being successfully challenged in court (you can't really rely on consent after the employment relationship begins).

This is opt-in by the end-user and disabled by default, so represents a genuine opportunity for employees and employers to decide if they want to gather this information and if so, what the opt-in/out procedure looks like.

This enhancement seems like a decent improvement on an existing feature. Folks are up in arms about a privacy problem which may exist in the underlying, pre-existing feature (updating your presence in the office in Teams/Outlook). This just simplifies the process for the employee to comply with whatever their employer mandates.

Unless, conspiracy hat on, they wanted to be prevented from serving Fable because it's too expensive for them to run, and they want some external authority to blame for shutting down access.

Is participation in your off time mandatory? If so, the problem is not hackathons. If not, why not just… not participate? The folks who do often enjoy working on the problems, networking, etc., so I’m not sure killing them all together is particularly fair to folks who do get a lot out of them.

Either these are already counted in the 9.5M, or they will continue not to be counted.

As far as I understand, action begins when the population hits 9.5M, so likely no-one gets kicked out, but fewer new visas will be approved, etc.

A way for computer programs to talk to each other on the same device as though they were running on different devices connected over a network.

I agree with you by the way, I just don’t think this is one of those cases.

That’s the most important thing. If we keep reading, maybe we can hold our own.

It’s not true that it’s not allowed in the EU. There’s the Barbalescu ruling which is case law that says employers must fulfill a bunch of criteria around informing employees, the necessity of the monitoring, and they are not allowed to impose blanket bans on private use, but it is still legal to monitor employees in the EU.

As of ~3 years ago (and according to this: https://www.onenewchange.com/en/roof-terrace#:~:text=Open%20... ) I think so yeah, but the restaurant/bar try really hard to make it look like you're not supposed to be there without actually telling you that you're not allowed to be there.