ajb

The UK used to have a distributed system - everyone had to have a solicitor store "deeds" of their property, which were a sort of paper blockchain of all the transactions the land had been in since - I don't know, since records began I guess. Since we got a centralised land registry cheaper solicitors have binned these, but some properties still have them as a historical record.

I understand the frustration, but it's justifiable for the public to be concerned about AI in general because the novelty of the technology means that boundaries between beneficial and problematic usage is not yet stable and well defined. For existing technologies consensus on that has usually been reached, although it changes over time. In many cases we may not yet have enough information to decide.

Nor is this a trivial decision. AI has the potential to change society and economic relations as profoundly as the industrial revolution, or the invention of the printing press.

These boundaries are also contested, as interests which benefit from a particular application are different from those whose interests are harmed. Society needs to identify which usages have a net benefit.

It also needs to define which usages cause "absolute" harms which is will consider unacceptable regardless of benefits to some parties. Such as, potentially, reductions in personal autonomy, increased leverage or dominance by government or private interests.

Not only that, but data and models which were collected/ built for one purpose can easily be adapted for others.

This is also basically now happening all at once in many domains.

In short, you can expect there to be tension over these boundaries for some time. It's not realistic to expect that others will agree with your personal perception of which applications are "obviously" unproblematic.

A positive one, or a negative one?

The problem with this line of thinking is that the more humanity succeeds at becoming efficient at providing for its essential goods and services, the more the economy rests on unnecessary ones. Should we really accept that the more successful we are, the less individuals can expect to be valued?

Yeah that happens all the time. Anyone/thing with popular public releases has fans/journeys scraping the website looking for unreleased material or scoops.

In the early days one of the high profile soaps in the UK published their "catch up" summaries for the week ahead which you could get just by editing the date in the URL. But back then not so many people were looking, so they were doing it for months...

In the semiconductor industry this role is called Field Application Engineer. They do serious work, not just slideware, as your chip probably ships with drivers that were designed months or years before you could get significant time from the customers engineers (which generally only happens after you manufacture, but you need software to prove it works well before that). So these guys are the ones who adapt it, and their feedback is valuable as they are the ones who build understanding of the customer.

However, at my old employer they didn't get commit rights to the main software repos. They had to carry around a bunch of patches which were gradually cleaned up and integrated. As I didn't directly work with them, I don't know if this says more about them or the guys managing the internal development.

It's a role that fits a different personality to that typical among software engineers. If you're bored as a dev it may suit. Pointless to try to shoehorn people into the role that don't suit it though.

Interesting, I hadn't heard of Petersen's algorithm before. There's a similar algorithm by Lamport (of course): https://lamport.azurewebsites.net/pubs/fast-mutex.pdf .

This was a while back, they have both now passed; but thanks for the good wishes.

An anecdote, but some reason to think that overworking part of your brain is a bad idea:

I had an Aunt who had dementia. This is obviously a terrible outlook. But her and my Uncle seemed to be doing okay. My Uncle was an ultra-competent guy, highly stable, the person you could rely on; and had been his whole adult life. So it was shocking when he had a mental breakdown and became manic. There was probably something physical going on, but what was also going on was that he had wanted my aunt to be able to live normally for as long as possible, so had been covering everything - for a year or more he'd had to be alert 7 days a week in case my aunt tried to cook on the gas stove or something like that, at which she was no longer safe. So the risk-alert part of his brain had been constantly overworked.

I appreciate that this is scientific research, but there are definitely companies out there that will try to row-hammer everyone's brain if this sort of thing is not heavily controlled.

Historically, vehicles actually did have completely interchangable motive components. They were called horses.

It would have been difficult given the state of other technology at the time for the inventors of the internal combustion engine to have supplied it as a drop-in horse replacement for your carriage, but you could kind of imagine that working with current technology.

It eventually worked for you - great. That doesn't prove that it will eventually work for everyone.

At some point, when something is not working, you have to write off the sunk cost. In many cases, there's no obvious threshold as to when you make that decision, and it's easy to think that maybe, if you'd just held out a little longer, been a bit more determined, you could have made it work. But you have to make a call, without knowing for certain either way. And it can be the right call, even if, with perfect information, you'd have made the other choice.

So, while I'm sure lots of the people saying "you should just have tried harder" are doing so with complete good will, it's not helpful. You don't know, for another random person on the internet:

* how much work they put in

* what other burdens they were carrying

* what other needs and responsibilities they could use that time and energy on

* how their mind is different to your mind

etc, etc. In short, you don't have most of the information needed to make that decision for them, and you don't have any call to judge them for it.

This isn't the worst article, and it's triggered a decent amount of discussion (despite being very short). However, I really dislike "What you're doing wrong/failing to do" titles. They are intended to trigger anxiety, which is manipulative and (in this case) precisely contradicts the concern the author is purporting to have for the rest of us.

On the subject: some people find meditation very helpful, others find it a net negative, or useless, or impossible to do. So a categorical "you should do this" isn't correct or particularly helpful. Try it, if it works for you, great; but don't put it about that people who aren't doing it are being negligent in some way.

This does violate the principle of least surprise. But,if you look through their other github issues, it becomes fairly clear that Anthropic are not interested in building an effective sandbox. It's a feelgood sandbox. To be fair, they must have a better focus on Claude not doing stupid things, or they would have crashed and burned by now. Nevertheless, if you want guaranteed limits, then they have to be enforced by external means.

Indeed. I wonder if it falls foul of labour law. Blacklisting is illegal and whitelisting (certification) is normally done with multiple competing third party certifiers.

Even if you're not a good photographer, wide angle lenses make rooms look enormous, by exaggerating the size difference between close and far objects. Before AI most estate agents used that.

In this case, this is probably not the only stereographic tattletale.

Had a competitor pull something like this with a previous employer. They were supposed to be interoperating with a standard, but they had a secret steganographic handshake, which they used to pretend that competitors products were unreliable (they had a first mover position in a smaller national market with specific requirements, so this wasn't shooting themselves in the foot). Our guys figured out the handshake and just silently implemented it. In this case, the competitor wasn't big enough to waste engineering time on multiple such hacks, but Anthropic have time (or Claude does).

That's interesting, because it's not my experience. A lot of the technical debt I see is that someone half-assed something thinking it would be easy to improve later, but the layer violations and inadequate tests make doing so a massive project, once it's become load-bearing.

They (or at least some of them) use waterfall - the real waterfall, not the bogeyman invented by agile consultants.

I think we can already conclude that the current approach is failing. I'm suggesting reasons why I think that is the case. Of course, my diagnosis may be wrong. But it's implausible that doing what we've done during 20 years of erosion of privacy is going to suddenly turn things around.

As for not alienating allies: if it's this easy to piss off people to agree with you, maybe it's also worth thinking about how to talk to those you want to win over.

The "open rights group" was supposed to be the UK EFF. My personal evaluation is that they haven't achieved much; YMMV.

Do you want to help stop that, or do you just want to feel smarter than other people? If you happened to have seen my comments in other threads, you'd know I'm against all this. But telling people they are idiots isn't going to win them over. This is not just a criticism of your comment, it's widespread in these discussions.

Opposing this requires:

* Linking to specific harms, which the public can emotionally resonate with. For example, scanning billions of photos for suspicions of child abuse will result in false positives that cause innocent people's kids to be taken away.

* Not seeming like an overheated conspiracy theorist. Feeling angry about this is legitimate, but it's not necessary to communicate in the same emotional register that you are feeling, even if it feels inauthentic not to. The public are saturated with people being publicly angry, much of which is purely performance. Deep concern may work better.

* Have plausible reasons for why this is happening. Yes, a few individuals like Thiel want to create a digital Stasi, but this would still be happening without them. Mostly this is driven by companies that want to make money, and officials who have a bias towards centralised processes, and are tunnel-visioned with respect to some issue. And people who are genuinely concerned about kids and haven't been given another convincing solution.

* Get facts straight. (Eg, rent/job IDs aren't a future threat. They are here)

Renting and work already require ID in the UK. Every landlord and employer is supposed to take a copy of original documents proving the right to rent/work in the UK. Technically you can do that without handing the docs to the government, but there's less potential liability to do so via the Home Office website.

In the case of coca cola, because use of coca leaves is highly regulated due to the fact that they also contain cocaine. There is a YouTuber who claims to have reverse engineered Coca-Cola, but he had to use tea-tree oil instead of actual coca leaf extract.

Here's EFF on reverse engineering and the law: https://www.eff.org/issues/coders/reverse-engineering-faq

Historically a lot of competition in physical products was very much reverse engineering. Because you can buy them without signing your rights away. That's why companies are keen on patents and click-through agreements.

If you look at how "clean room" processes work, they are actually a form of reverse engineering. Also clean room technique exists to avoid your new implementation infringing copyright, not trade secrets.

For something to be a trade secret, you have to actually keep it secret. If I get the ingredients of Coca-cola from an ex-employee, I've stolen a trade secret. If I work it out by doing a chemical analysis, I've stolen nothing.

There is a difference with anthropic, as no-one signs a licence agreement to buy a coke. But Anthropic are also not saying you can't publish the output of their models. It's not clear to me if trade secret law will (or should) cover a secret which can be extracted from information that licensees are not restricted from publishing.

In a way, the cack-handed way they've gone about this makes me slightly more optimistic. If we must have such a law, please let it be one which:

* Creates a market for privacy tech of several million teenagers

* Wastes police time chasing down social forums which kids are hosting abroad using their pocket money

* Rubs the noses of the securirati in the fact that they've made it easier for terrorists to hide their comms among the thousands of teenage speakeasies

This is not the 80's when comms tech required capital and man-years of engineering. Setting up forums online isn't even a high-school project.

The cops should know that trick by now - the "distraction car" is literally the plot of "Smokey and the Bandit" from 50 years ago. Nevertheless, you may be right :-)

Yeah sounds like it's backwards , and should be "in order to be a member of the customer club, it is a condition to receive marketing / offers ."

Ah, but that is the old world, before AI. Given AI, dumb leaders can trash a business at the speed of thought.

However, you can get many of the benefits of a "local model" by outsourcing all the hardware maintenance but still using an open model. Guaranteed repeatability for one.

A lot of the reason people outsource normal software is its brittle security properties, not sure that even applies to an LLM - it can go and look up the latest security best practices just like an engineer can.

There's a big difference between wanting a succinct proof and wanting a succinct statement of the requirements. The easier it is to state the requirements, the more likely you have stated them correctly. Whereas succinct proof is not relevant for industrial purposes, as long as the proover has a small trusted kernel, it makes no difference to the reliability.