Is it fair to characterize a founder of a failed startup as a con man? I.e. did he make claims that were factually untrue, and intentionally deceived investors?

Disclaimer: Subjective opinion, as a founder.

It’s as fair as characterizing an athlete saying he’s competing to win.

A common logical/semantic mistake is that claims about future facts are lies. They can be exaggerations, and we do see this often and I’m personally annoyed when I see it. These claims can be genuine or not. I often think it’s easier for a founder to be naive (or crazy) than not, because they can believe (sometimes strongly) the things they say.

But neither are lies. A lie is when you misrepresent a fact, something that already happened. That’s a big causal difference.

The large gray area is when someone misrepresents intent (as with most cases for cons). This, I suspect, is the main questions you should be asking, but I suspect it falls into the ethics rather than legal debates.

Yes! He did, so it is.

> There is no promise of delivered product

There absolutely is a promise. Even if you manage to legally find a way to not get sued, taking advantage of the fact that everyone who gave you money believed it was a promise is still scamming them.

Isn't the whole thing (or two) with Kickstarter is that if it's not funded, everyone gets their money back and if it's funded, the creator tries to deliver the goals according to the timeline but if they don't, they're not held liable for that? So if for some reason the creator run out of money before they could send actual products, you as a donator don't have the right to get your money back? Maybe I misunderstood the whole concept of Kickstarter.

James Proud:

* Promised an alarm clock that would do a bunch of things

* Took $2.5M in funding from Kickstarter

* Took another $50M in funding from elsewhere

* Delivered a piece of hardware that did essentially none of what was promised.

It's all detailed in OP and the linked Verge article. That's a scam and I'm not interested in your legalese arguing whether they can be sued or not.

> I'm not interested in your legalese arguing whether they can be sued or not

What... You know, it doesn't matter. Thanks for the summary anyways!

Ethics and legality are independent concepts. A scam is an ethical construct.

>"Even if you manage to legally find a way to not get sued"

What is the counterfactual? Without knowing the number of attacks prevented by these tools, we don't know what the baseline would be.

For the record: they prevented essentially nothing in our muni. We're 4.5 square miles sandwiched between the Austin neighborhood of Chicago (our neighbor to the east; many know it by its reputation) one side and Maywood/Broadview/Melrose Park on the other, directly off I-290; the broader geographic area we're in is high crime.

We ran a pilot with the cameras in hot spots (the entrances to the village from I-290, etc).

Just on stolen cars alone, roughly half the flags our PD reacted to turned out to be bogus. In Illinois, Flock runs off the Illinois LEADS database (the "hotlist"). As it turns out: LEADS is stale as fuck: cars are listed stolen in LEADS long after they're returned. And, of course, the demography of owners of stolen cars is sharply biased towards Black and Latino owners (statistically, they live in poorer, higher-crime areas), which meant that Flock was consistently requesting the our PD pull over innocent Black drivers.

We recently kicked Flock out (again: I'm not thrilled about this; long story) over the objections of our PD (who wanted to keep the cameras as essentially a better form of closed-circuit investigatory cameras; they'd essentially stopped responding to Flock alerts over a year ago). In making a case for the cameras, our PD was unable to present a single compelling case of the cameras making a difference for us. What they did manage to do was enforce a bunch of failure-to-appear warrants for neighboring munis; mostly, what Flock did to our PD was turn them into debt collectors.

Whatever else you think about the importance of people showing up to court for their speeding tickets, this wasn't a good use our sworn officers' time.

> As it turns out: LEADS is stale as fuck: cars are listed stolen in LEADS long after they're returned.

Is this related to rental companies reporting cars as "stolen" if they are an hour overdue on their scheduled return?

Can you elaborate on why you're not thrilled about Flock being removed?

The metro area is blanketed in ALPRs and we were the only ones actually writing real policy about them. Now we don't have any ALPRs and can't build policy or shop it to any of our neighbors. We had harm reduction for the cameras and a plausible strategy for reducing their harm throughout the area, and instead we did something performative.

Why is it better to reduce the harm of a practically useless anti-crime device than remove it entirely?

That's a good and reasonable question. The answer is: the cameras weren't going to do any meaningful harm in Oak Park (they were heavily restricted by policies we wrote about them, and we have an exceptionally trustworthy police department and an extremely police-skeptical political majority). But you can drive through Oak Park in about 5 minutes on surface streets, and on either side of that drive you'll be in places that are blanketed with ALPRs with absolutely no policy or restrictions whatsoever.

Had we kept the cameras, we'd have some political capital to get our neighboring munis (and like-minded munis in Chicagoland like Schaumberg) to take our ordinances and general orders as models. Now we don't. We're not any safer: our actions don't meaningfully change our residents exposure to ALPRs (and our residents weren't the targets anyways; people transiting through Oak Park were) because of their prevalence outside our borders.

What people don't get about this is that a lot of normal, reasonable people see these cameras as a very good thing. You can be upset about that or you can work with it to accomplish real goals. We got upset about it.

> What people don't get about this is that a lot of normal, reasonable people see these cameras as a very good thing. You can be upset about that or you can work with it to accomplish real goals. We got upset about it.

An alternative is you can try to convince those people that, while their desire to reduce crime is perfectly understandable, this might not be the way to do it effectively, to say nothing of the potential avenues for abuse (and in current day America, I'd be very wary of such avenues)

It remains an issue of trust for me. You not only have to trust your police and government(s), but you have to trust Flock too - and that trust has to remain throughout changing governments and owners of that company. I have a healthy distrust of both, particularly lately.

Just as importantly, but more to the point, is still the question of whether they're actually useful. To that end, does not the same logic apply to being able to pressure nearby municipalities to remove the cameras?

In any case, while I remain fundamentally opposed to such surveillance, you raise very good points, and I appreciate you taking the time to explain your position in this thread.

I'm fine that we took the cameras down. As you can see from my first comment on this part of the thread: they weren't working, and before we stopped our PD from responding to stolen car alerts, they were actively doing harm. But I disagree with you about the long-term strategy. I'd have kept the cameras --- locked down (we had an offer from Flock to simply disable them while leaving them up, so that they wouldn't even be powered up) --- and written a formal ALPR ordinance. Then I'd have worked with the Metro Mayors Caucus and informal west suburban mayor networking to get other munis to adopt it.

Why do you have political capital to convince neighboring counties to copy your legislation, but not to copy your decision to remove the cameras?

(And you can still pass legislation restricting cameras even when they aren’t in your county…)

Because regulating cameras is an easier sell than disabling them, because ordinary people do not share HN's priors about surveillance technology, like, at all.

I'm not sure I agree with this, unless by "ordinary people" you mean a particular group. In my experience, the vast majority of members of oppressed or marginalized groups are strongly against these things. The only people I know who defend it are those that can hide behind "if I'm not doing anything wrong what do I have to hide"

It's not like we're going to have to argue about this in the abstract for long. Check in with me in a year and we'll see: are there more ALPRs or less ALPRs in the near west suburbs of Chicago? I would put money on a lot more. Dozens more were getting rolled out in our neighboring munis right as we announced we were canceling our contract.

I'm sure you're right, but that doesn't imply to me that "ordinary people" are okay with surveillance technology. At least 1 other explanation would be that they don't understand the implications.

Anyway, we'll check back in a year and see: are they actually effective and used responsibly? I would put money on "no"

I'm not wondering about it; I've watched the two sides of this issue play out. Progressive activists turn out with stuff about surveillance and resisting the Trump administration (I'm sympathetic!) and liberals (we only have progressives and liberals here) turn out talking about public safety and how the camera enforcement is less racist than human discretionary enforcement (I'm sympathetic to that argument too, but as I've noted elsewhere on the thread: the cameras aren't effective here).

This played out over years here; I attended all the board meetings, transcribed them, took notes, kept tallies of who was saying what. I was involved in our last election and the two mayoral candidates squared off on this issue (among several others).

We also don't know the number of attacks indirectly caused by these tools, by instilling a more fraught social environment.

I don't care. The world is a dangerous place, we make it safer by promoting freedom and education and goodwill and faith in people, not by growing the police state. We do know for a fact however that in the near future anything "think of the children" or "just looking for criminals" ultimately gets turned against all of us as the government grows and grows without limit, our rights will become fewer and fewer with the encroachment. It's not "panic" or "exaggeration" it has happened all through history of nation-states.

Not a chance. Even if American companies did abide by it, there is no reason Chinese companies would. And good luck definitely proving that a model trained on it.

That problem along with its many solutions are surely littered throughout the training data. Not to mention, it would be trivial to overfit on that problem. I don't know why people still reference that.

> That problem along with its many solutions are surely littered throughout the training data. Not to mention, it would be trivial to overfit on that problem.

It would be trivial to over-fit, if that was their goal.

But why would there be a large number of good SVG images of pelicans on bikes? Especially relative to all the things we actually want them to generalise over?

Surely most of the SVG images of pelicans on bikes are, right now, going to be "look at this rubbish AI output"? (Which may or may not be followed by a comment linking to that artist who got humans to draw bikes and oh boy were those humans wildly bad at drawing bikes, so an AI learning to draw SVGs from those bitmap pictures would likely also still suck…)

Because it's become the iconic test for them and countless articles have been written about it with plenty of examples.

I added the word "good" in there, you may have replied before seeing that edit.

Maybe we can try “dog in a paraglider”? If it fails then we know it’s overfitting, if it works then the model generalises well?

Honestly, you're probably right. It's quickly become a pretty weak eval, but the guy that's running that eval is excellent. I'd much rather the evals people were using to test these things looked more like classic/boring engineering problems: deploy to dev/test/stage/prod with digital ocean, cloudflare, github, and a common git flow. Boring problem, I know, but that problem is wildly complex when you start to add a few extra dimensions (frontend vs backend, ports shifting between deployments, local deployments, etc.).

i think the point is people assume models arent overfitting for it, and its a fun/silly way to potentially gauge its general abilities

It's a result of the system prompt, not the base model itself. Arguably, this just demonstrates that the model is very steerable, which is a good thing.

It wasn't not a result of system prompt. When you fine tune a model on a large corpus of right-leaning text don't be surprised when neo-nazi tendencies inevitably emerge.

It was though. Xai publishes their system prompts, and here's the commit that fixed it (a one line removal): https://github.com/xai-org/grok-prompts/commit/c5de4a14feb50...

If that one sentence in the system prompt is all it takes to steer a model into a complete white supremacy meltdown at the drop of a hat, I think that's a problem with the model!

The system prompt that Grok 4 uses added that line back. https://x.com/elder_plinius/status/1943171871400194231

Weird, the post and comments load for me before switching to "Unable to load page."

Disable JavaScript or log into GitHub

It still hasn't been turned back on, and that repo is provided by xAI themselves, so you need to trust that they're being honest with the situation.

The timing in relation to the Grok 4 launch is highly suspect. It seems much more like a publicity stunt. (Any news is good news?)

But, besides that, if that prompt change unleashed the very extreme Hitler-tweeting and arguably worse horrors (it wasn't all "haha, I'm mechahitler"), it's a definite sign of some really bizarre fine tuning on the model itself.

What a silly assumption in that prompt:

> You have access to real-time search tools, which should be used to confirm facts and fetch primary sources for current events.

xAI claims to publish their system prompts.

I don’t recall where they published the bit of prompt that kept bringing up “white genocide” in South Africa at inopportune times.

These disgruntled employee defenses aren't valid, IMO.

I remember when Ring, for years, including after being bought by Meta, had huge issues with employee stalking. Every employee had access to every camera. It happened multiple times, or, at least, to our knowledge.

But that's not a people problem, that's a technology problem. This is what happens when you store and transit video over the internet and centralize it, unencrypted. This is what happens when you have piss-poor permission control.

What I mean is, it says a lot about the product if "disgruntled employees" are able to sabotage it. You're a user, presumably paying - you should care about that. Because, if we all wait around for the day humans magically start acting good all the time, we'll be waiting for the heat death of the universe.

or pr department getting creative with using dog whistling for buzz

Where is xAI’s public apology, assurances this won’t happen again, etc.?

Musk seems mildly amused by the whole thing, not appalled or livid (as any normal leader would be).

Is it good that a model is steerable? Odd word choice. A highly steerable model seems like a dangerous and potent tool for misinformation. Kinda evil really, the opposite of good.

Yes, we should instead blindly trust AI companies to decide what's true for us.

Doesn't matter. In fact, it makes it even funnier: all these investors spending billions of dollars on OpenAI just end up subsidizing the competing models.

That was trained on data scrapped from the web. I'd say it's fair.

It would happen in China regardless what is done here. Removing billionaires does not fix this. The ship has sailed.

The benchmarks agree as well.

Isn't that how previous models were, before the attention is all you need paper?

Or be in the business of building infrastructure for AI inference.

Is this not the same argument? There are like 20 startups and cloud providers all focused on AI inference. I'd think application layer receives the most value accretion in the next 10 years vs AI inference. Curious what others think

Or be in the business of selling .ai domain names.