openasocket

I actually really disagree with this. If we want our representatives to reflect the will of the people, they need to have an accurate understanding of their electorate. When issues come up we need a way for the people to be heard, and polling is an important part of that.

I’m not necessarily disputing your point, but it’s very odd to downplay one source for potential bias, and then follow up with a study from the CATO Institute. Like, the CATO institute is openly a libertarian think tank founded by one of the Koch brothers.

I also think it’s unfair to look at just food prices. Housing and healthcare have bloomed massively.

I try to do this as well. It certainly feels better and makes me more confident that the code will make sense when it implements it. But I’ve found even with a very firm design an LLM will undermine it in the details. It will be convinced there is a particular way to do something, I will tell it not to do that and give it a complete alternative, and it will say it implemented it as written. And if you just look at the data structures and function signatures things look fine. But you look at the actual implementation details and it will try in some awkward way to implement things totally differently.

She was acquitted of several counts of wire fraud to defraud patients, but that does not mean harm was not found to customers. I haven’t been able to find the specific statute, but essentially the jury found that the prosecution didn’t prove all of the elements of that charge beyond a reasonable doubt. We’ll likely never know the details, jury deliberations are secret. But there was absolutely harm to patients. Balwani was also charged with patient fraud and WAS convicted on those counts.

But the question was how is this deceitful? Like, when someone says a documentary is deceitful, I assume there’s some important information the documentary didn’t include, or provided outright false information. I get that the documentary went in a direction you didn’t like. I just saw it this afternoon, and didn’t feel the same way, but I’m still processing it and I do find your opinion interesting. But unless you are saying this is, like, false advertising, I don’t see how that qualifies as deceitful.

It’s been a while since I studied complexity theory, but is that what Blums speed up theorem says? A quick google search suggests it states that there exist computable functions where there is no optimal program that computes it.

Like, there exist problems that have a lower bound of n time where algorithms can get arbitrarily close to O(n) complexity but there is no program that actually reaches it (there is one that runs in time O(n^{1.1}), one that runs in O(n^{1.01}), etc).

But that doesn’t mean that there aren’t hard walls, it just means there are hard walls where every program’s runtime is > that wall, rather than >= that wall. But maybe I’m misunderstanding, or there are some corollaries I’m missing?

As an aside, as someone who studied more on the pure math side, Blums speed up theorem is rather intuitive. It looks like there might be a relatively elegant proof using topology. Blums speed up theorem is basically just a statement that the set of programs that implement a particular computable function doesn’t satisfy the greatest lower bound property. This is also the case for, for example, the set of rational numbers: there exist subsets of the rational numbers with a lower bound but not a greatest lower bound (like the set of all positive rational numbers q where q^2 >= 2). I’m hand waving, but I think in that same vein you can show that the set of programs under the topology of ordering by runtime is completely disconnected.

This is true for the general criminal conspiracy law, but be aware that an “overt act” doesn’t have to be an illegal act, just some action in furtherance of that crime. That can be purchasing a weapon, or scouting out a location. There may also be other laws in play depending on the specifics

Nukes are arguably ONLY useful in defense (and deterring another from attacking). Using nuclear weapons against an opponent that also has nuclear weapons will very likely lead to a nuclear exchange and the decimation of both countries. It’s suicide, and so no one is going to do it unless they have nothing left to lose. A nation under attack, facing an existential threat, and about to be overrun, is in exactly that circumstance.

And yeah, maybe the defending countries leader isn’t willing to end it all, and maybe try collaboration or partisan resistance. But if you were the attacking country, would you be willing to make that bet? Or would you hesitate?

A defense pact doesn’t always hold, though. Poland had a mutual defense pact with France and the UK in September 1939. And even if China were to aide the DPRK in a crisis, the DPRK would be the junior partner in that arrangement. In this scenario, if China is providing the lions share of manpower and weapons, they get to decide strategy, and dictate terms for war termination.

The DPRKs primary goal is preservation of the Kim regime above all else. That’s why the DPRK over the last 70+ years has focused on Juche, a principle of self-reliance. The vast majority of their weapon systems are of indigenous designs and manufacture. And why they have developed and deployed an independent nuclear deterrent.

I think you have to make a distinction between bugs and actual vulnerabilities. Therac-25 killed people, but I wouldn’t consider anything about it to be a security vulnerability. In my mind, the distinction between a bug and a vulnerability is that a bug is triggered during “normal” operations and can do anything. Whereas a vulnerability requires an adversary to “trigger” the vulnerability, and can do this to achieve some cognizable malicious goal. There’s probably some overlap on the edges; whether an issue in a library is a bug or a vulnerability may depend on how it is used, for example.

But I think the most important thing to keep in mind is that a bug isn’t necessarily less serious or less important than a vulnerability. A serious bug should be patched just as urgently as a serious vulnerability.

Not a lawyer but I have some knowledge of criminal law in general. There are a lot of details left out, but I do think she is likely to get a decent settlement. Though the Flock/AI part of it isn’t really the reason.

So, when you get arrested, you are required to be arraigned by a judge, and that judge has to find probable cause for arrest. That requires the prosecution to present evidence to the judge that a crime was committed and you probably did it. It’s a very low standard of proof, but it does have to happen. And then you are required to get bail, outside of very extreme cases. It depends on the jurisdiction, but all of that has to happen relatively quickly after arrest, like 24-72 hours. If she didn’t get bail until 13 days after the arrest, that’s a pretty cut and dry civil rights violation. If they didn’t get a probable cause determination in that time period, that’s a civil rights violation (unless they got an arrest warrant prior, which they might have in this case).

If she did get bail within a couple days, but it was just so much she couldn’t afford it, then it gets murkier. Because at that point thats not the police or the prosecutor, that’s the judge. Similarly, if the prosecutor presented the evidence they had to a judge for a probable cause determination (or the arrest warrant) and the judge granted it, that’s on the judge. You would have to show the prosecutor and/or police lied or presented inaccurate information in court. If not… at that point you have two independent branches of government screwing up.

Wait, was she not arraigned for 13 days? The article makes it sound like it wasn’t until 13 days after the arrest, and her lawyer presented evidence, that she was let out. That… that can’t be right. It varies from state to state, but generally you are required to be arraigned within 24-72 hours of arrest. And then they are required to give you bond (a no bond hold can happen but is very rare and I don’t think any jurisdiction allows that outside of murder).

Was she just given an extremely high bond she couldn’t afford? Not that that’s much better than no bond at all, but I’d really like to understand the details here.

One thing I find surprising is everyone is talking about the danger of an agent going rogue but not the danger of an agent getting hijacked. These companies are making this clusters with thousands of agents running at once, with frontier, often not-yet-released, quality models and massive computational and network resources. And these things are given access to whatever they want on the Internet. Even if that was restricted to read-only access to the Internet, that’s still exposing the agents to untrusted input. All it takes is some bad actor creating a website that attracts one of these agent swarms and doing prompt injection. Then your fancy AI cluster will start doing whatever that attacker wants. And the fact that we have multiple examples of these swarms trying to coordinate on random corners of the internet shows they are almost pre-disposed to it.

Now it feels like companies are treating these breakouts like a chance for PR. I don’t think that will change until their swarm gets corrupted by some random black hat to do en-masse spear phishing or something

A lot of military comms are point-to-point or broadcast between radios where TCP/IP doesn’t really make sense. On top of that, you need to understand how much legacy military networking stuff is out there. This stuff long predates the internet. We had tactical data links and target tracking where a ground station could receive a target info from a remote radar array, and automatically direct fighters on autopilot on an intercept course with enemy contacts in like 1960 (see SAGE as an example). These systems have been around for decades and been continually improved on. There’s a lot of inertia to maintain compatibility. Because you really don’t want someone to die because your ally tried to warn you but their tactical data link isn’t compatible with yours.

This is actually not true! This is a pedantic point but I think is important.

The department of defense is not equivalent to the pre-1947 War Department. The War Department was in charge of just the army. There was a separate Department of the Navy. There was no over-arching department for all of the US military.

Post WW2 the US military was restructured to organize all of the branches (including the newly-formed Department of the Air Force) under a united civilian leadership. This was done in the National Security Act of 1947. The new structure was initially called the National Military Establishment and then renamed to the Department of Defense.

So the department of war name going away wasn’t a PR thing. It was because of military parochialism. If they had called this unified structure the Department of War it would have seemed like the Army was now taking over the other branches! And I promise you today’s Admirals understand that history, even if the secretary of defense doesn’t.

Ok, so the article buries a lot of the context here. To clarify: plaintiffs can 100% sue the city for this under tort law (I believe the clean water act, but there’s probably other causes of action). This particular case was about suing the city officials in their personal capacity. So making those individuals pay, instead of/in addition to the city. Government officials are generally protected under qualified immunity. One exception to qualified immunity is 42 USC 1983: deprivation of rights under color of law. But that requires them to prove that these officials violated some constitutional right, beyond just some statute.

The victims can sue the city/state. They might already be doing so in parallel. And honestly that’s the only way to be made whole: these city officials don’t have the tens of millions that would be required to compensate victims. This is more about holding government officials individually accountable.

They absolutely can, depending on the jurisdiction. I saw plenty of cases out of Atlanta with drug possession cases with no prior history getting an ankle monitor, back when they were live streaming bond hearings from Rice street.

I’ve always wondered why candidates need so much money. I mean, this doesn’t go directly into their pockets, it’s largely used to fund ads, right?

I have a pet theory about political ads that I would love to find data to be able to confirm/deny. My theory is that those political ads don’t actually convince anyone about who to vote for, and aren’t intended to. Instead, it’s about convincing the people who already support you and your party to show up and vote. If that’s true, then the root cause of all of this money being injected into politics is really a symptom of low voter turnout. And so anything we can do to improve voter turnout will lessen the influence of money in politics.

My justification for this theory is that I see all of these political ads and really don’t find them convincing or liable to change my vote. I actually find attack ads often have the opposite effect on me, where I am immediately skeptical of whatever claim they advance and feel sure there’s important context they are leaving out. Of course I know I’m not the average voter, and other people may see it differently, but I’d love to see some data on this.

Sure, but as I said in my second example, forcing oil and natural gas companies to prevent methane leaks would have much more of an impact than reducing air travel, and requires no sacrifice for the average person. When considering which policies to advance, we should look at both the potential impact AND the democratic support for it.

Climate change deniers and polluters want to paint the environmental movement as a bunch of neo-Puritans that deride anything that isn’t the bare minimum of subsistence as immoral. And I worry they’ve done this for so long that some supporters of the environment are starting to believe the lie.

Short answer - yes. If you post bond, and you don't violate the terms of your bond, you get your bond back when your case is resolved, whether you were found guilty or not guilty.

If you do something that violates the terms of your bond, the prosecution can file a motion alleging a violation, at which point the defendedant can admit to the violation or deny and request a hearing. Then there's an evidentiary hearing, basically a mini-trial, where the judge hears evidence and determines if there was a violation. If that happens, the consequences can depend. You can forfeit your bond, your bond can be raised, new bond conditions can be imposed, etc.

If you were given a cash bond and can't afford it, you can contract with a bail bondsman. That bail bondsman will pay the bail for you, and in some sense takes responsibility for you not violating your bond, in exchange for some fee (usually 10% of the bond amount). That amount is NOT refunded

For those that don't know, grand juries declining an indictment is extremely rare. A grand jury is basically a check on prosecution, that they have to have some initial evidence before charging someone with a felony. The standards are much lower than the subsequent criminal proceedings.

The grand jury only hears from the prosecution, there is no defense involved. Only a majority of the grand jury has to sign off, not unanimous like in an actual trial. The standard of evidence is just probable cause, not beyond a reasonable doubt. The rules of evidence are relaxed, meaning hearsay and other evidence can potentially be introduced that would normally be barred from a trial.

Because of the above, the rate of indictment from a grand jury is very high, over 90%. Most prosecutors will go their entire careers without getting a "no true bill" (meaning the grand jury did not sign off on an indictment). There's a saying that "a grand jury would indict a ham sandwich." So the fact that there was no indictment here is a big deal. It will probably hurt that prosecutor's career.

It’s an effective a surprisingly old technique, being used on cruise missiles as early as the 1960s. It actually precedes GPS and satellite navigation by several decades. Im continuously blown away by what engineers were able to do in that era with such limited computing power. Take a look at SAGE, for example.

Fun fact: the usage of TERCOM in the tomahawk missile actually limited its ability to be used in Operation Desert Storm. Routes had to be planned to go around actual topographical features, instead of hundreds of miles of flat desert.

Important to note that this was not a decision on the merits. The court didn’t find that the decision was written by AI; just that the facts alleged would be covered by judicial immunity. Proper action is to appeal, and file a grievance with the Nevada Council of Judicial Discipline, possibly also the Nevada Bar Association. Those institutions take these things seriously, and the judge can end up with her license suspended or even disbarred. A license suspension, even a short one, is very serious for a judge. A judge is forbidden from having any other employment, so a judge under suspension isn’t getting a salary and can’t get other employment. So it’s a pretty significant monetary penalty.

EDIT: Yeah this plaintiff is nuts. His complaint, linked by a response to my comment and in other places, is just bizarre. I would highly doubt any allegations he might make.This guy has apparently been suing a bunch of judges in the family courts in Las Vegas. A lot of those proceedings appear to be online: https://www.youtube.com/playlist?list=PLsKMpjAPcB2vRRCMtdEb1...

Ben Gvir is the minister of national security and has been for years. He literally is a representative of Israel. He may not represent the views of Israelis, but he is in power, and is using that power. That makes him a relevant example, as relevant as the day is long.

Whether something is “thinking” or not is really more of a philosophical question. It really depends on which of the many, often contradictory, definitions of “thinking” you choose. Sometimes we use “thinking” to describe advanced calculation or analysis, which would cover LLMs along with chess engines and many other algorithms. Other times we use “thinking” to describe what conscious beings (which is ALSO a philosophical term with many different interpretations) do, and I think most people would agree LLMs aren’t conscious. And then there’s a whole spectrum in between. We’ll probably need to come up with a whole new set of terms to describe the new and evolving capabilities of LLMs.

But for me, for any stronger definition of “thinking,” I don’t think the output of any LLM would actually convince me. Producing a result isn’t thinking - for all you know it is just printing verbatim something from the training data. No, to conclude if it is thinking or not I would want to look inside its head, at the architecture and watch it produce those results. And because LLMs are so different it will probably take advancements in mathematics or computer science to be able to really interpret what is going on

I know exactly what you mean. Something about those AI explanations just make my eyes glaze over. Dozens of new terms and metaphors and analogies conjured out of the ether to explain even the simplest thing. And when I try making it explain with examples, or show me the code it is proposing, often it seems unrelated or even in tension with whatever it tried to say before. I’ve given up trying to assign any meaning to those weird little soliloquy’s. I’m convinced that those don’t really have any meaning under them, and when you have it actually make a code change it does the actual work.

For those that aren't aware, the uncertainty principle goes pretty deep. You can actually define the uncertainty principle as an inequality involving the integral of a function vs the function's Fourier transform: https://en.wikipedia.org/wiki/Uncertainty_principle#Harmonic... . And in a quantum system you can construct the momentum of a particle as the Fourier transform of the position (up to a constant) and the uncertainty principle falls out because of this. This article doesn't state that super explicitly. So the real interesting thing being done here is getting a Fourier transform that works on fractal spaces.

With the expiration of New Start earlier this year, there are no nuclear arms control agreements in effect between the US and Russia, which hasn’t been the case since the 1960s. Russia, China, and the US are all actively modernizing and developing new nuclear capabilities. China has been greatly expanding their nuclear arsenal, building hundreds of new silos. North Korea has an expansive set of tactical and strategic nuclear systems and are in strategic cooperation with Russia. Russia is 4 years into a destabilizing land war, have developed several new weapons systems, and are now in a nuclear weapon sharing agreement with Belarus. And I’m sure I’m missing a bunch of other things.

We are in a period of strategic instability and an arms race that is unlike anything we have seen in the last 50 years. You can quibble with the specifics all you want, but I don’t think anyone can argue that we are at a lesser risk of nuclear war that we were, say, 20 or 30 years ago

Biggest thing I can recommend is to focus on the fundamentals, even if it doesn’t seem like it’s something you would use in practice.

Learn data structures and algorithms. Understand finite state machines and context free grammars. Learn at least some mathematics. Discrete math, Boolean logic, and statistics. Learn how operating systems work, networking, databases, compilers, machine code, low-level details like registers and the cache hierarchy. Learn more obscure programming languages: like Lisp, Prolog, Haskell, and Forth.

You probably won’t use those in your day job, but they will open your mind to fundamentally different ways of thinking about problems. Honestly, just try to learn as much as you can about as many different areas of computer science as possible. Don’t worry about learning version control, the latest web frameworks, or working with AI, at least not until the very end.

You want to come out of college with two things: the ability to learn new things quickly, and having a wide breadth of knowledge to draw upon. As someone who regularly interviews candidates, that’s what I’m looking for more than anything else.

+10, I can't stand AI-generated comments. Half the time they are paragraph-length translations of every 5 lines of code into English prose, or just regurgitating whatever prompt they were given. When they actually try making high level comments they are worse than useless. It's really hard to put into words exactly why. It's something about how they are very verbose, but at the same time vague and stated in a slightly-off manner. Or they will introduce new terminology not reflected in the code's nomenclature, sometimes terms that are just synonyms for terms that the code already uses. Something about it makes me instantly distrustful of any information the comment may actually convey.