"A military operation is being planned" is very different from "Maduro will be kidnapped in the next x hours".
"Pentagon planning a military operation" is not exactly classified information as it is safe to assume that Pentagon is always planning a military operation.
did anyone have any reason to believe that was classified information that was leaked, instead of just a random person speculating? if not, then he had no intent to leak that information. If a random soldier told you, "iran will be nuked tomorrow" do you believe them? especially on a speculation platform, for all you know he's also guessing based on the same activites and events the public is observing. laws are all about intent and state of mind, what actually happened is irrelevant, what was intended is what matters. For example, killing a person is not a crime in and of itself, if it was all soldiers who kill someone in combat would be in prison, as would people who kill in self-defense. Matter of fact, if no classified information was actually leaked, but it could be proved that he intended to do something to leak classified info (which requires others to believe it is truthful information, instead of speculation) then that in itself is a crime.
Saying anything at all on a speculation platform, especially if others don't even know your identity (or you have no reason to believe they do), can only be treated as speculative intent, not intent to disclose classified information.
If a random soldier bet me $32,000 that iran would be nuked tomorrow, I would believe them a lot more.
If you are a potential assassination target and you notice that a prediction market about your assassination has a sudden weird spike on a specific week, then you would likely take extra precautions in that week. After this incident, surely other world leaders and public figures are watching the prediction markets for exactly this.
The stated point of prediction markets is to aggregate private opinions into public predictions of the future. If you participate using classified information and influence those public predictions, then that's leaking that classified information. This is more true if the bets you make are large and the market is relatively small, then you will send a much more clear signal.
And literally every other thing they do on the internet.. remember that Strava shit? You have relatively technically unsophisticated people with high level access and not a lot of adult supervision. That seems like a juicy target. I assume there are a lot of well funded and staffed outifts around the world who have noticed the same thing.
There have been some cases where fitness tracker data shows where some military installations are located. Or when they're jogging on a ship that's taking them to deployment somewhere.
The Ukraine war has shown that cheap intelligence tricks can be used against the average recruit, like pretending to be a dating website and getting the GPS locations of horny enemy soldiers so your drones can drop grenades on them.
It doesn't need to be crypto wallet tracking. The amount of spyware being built into phone apps is where those agencies would be putting some effort into obtaining access to.
> "A military operation is being planned" is very different from "Maduro will be kidnapped in the next x hours".
IIRC, the bet was on "Nicolas Maduro out?":
> If Nicolás Maduro leaves office before February 1, 2026, then the market resolves to Yes.
So the bet wasn't specifically "Nicolas Maduro kidnapped?" or even "Nicolas Maduro out by January 3rd?" And IIRC there was a lot of Trump saber rattling about Venezuela in the days before, hence the creation of the bet. I could absolutely see a plausible way to link these publicly-available pieces of information into a winning bet:
Sure, but Netflix is not interested in whether you love Stargate or not. Telemetry says that you never click it, so it's ok to remove it from their catalogue (which is correct).
Now, they could've done a better job by increasing the quality, but that's a further (and costly) optimisation.
Netflix should be very interested in a fact like "Netflix has show X, but Netflix subscribers who love show X choose to watch it someplace else due to issues with Netflix".
That's how all time highs work, yes. And it didn't fall to 1980s levels, 2020 was tied for second highest all time in math and was 1 point behind second highest all time in reading.
> in retrospect it was too angry since he had no choice
Imho you did the right thing. Everyone has a choice.
I did the same (but instead of an email it was an angry speech, face to face), with the exact same question (if they would prefer my sending sick children to school). My child's grades were fine, too. Haven't received any letters after that.
> Micromanaging the model doesn't seem like a great idea when doing real professional work with professional goals/deadlines/pressures.
Remember that it's not only the cost per token, but also speed. Some tasks are done faster with simpler/less-thinking models, so it might actually make sense to micromanage the model when you have deadlines.
Haiku is most definitely not fine for the code bases that I work on. Sonnet is probably fine for most daily tasks, but Opus is still needed to find that pesky bug you've been chasing, or to thoroughly review your PR.
> Haiku is most definitely not fine for the code bases that I work on. Sonnet is probably fine for most daily tasks, but Opus is still needed to find that pesky bug you've been chasing, or to thoroughly review your PR.
Yeah, I hear that a lot, but it never comes with proof. Everyone is special.
I’m sure you’d find that Haiku is pretty functional if there were a constraint on your use.
I use models from Opus through Haiku and down into Qwen locally hosted models.
I don't know how anyone could believe that Haiku is useful for most engineering tasks. I often try to have it take on small tasks in the codebase with well defined boundaries to try to conserve my plan limits, but half the time I end up disappointed and feeling like I wasted more time than I should have.
The differences between the models is vast. I'm not even sure how you could conclude that Haiku is usable for most work, unless you have a very different type of workload than what I work on.
More information required. What are you working on? What languages? How do you define “small tasks”? What are “well-defined boundaries”? What is your workflow?
Most importantly, define your acceptance criteria. What do you mean by “disappointed” - this word is doing most of the heavy lifting in your anecdote. (i.e. I know plenty of coders who are “disappointed” by any code that they didn’t personally write, and become reflexively snobby about LLM code quality. Not saying that’s you, but I can’t rule it out, either.)
The models are not the same, but Haiku is definitely not useless, and without a lot more detail, I just ignore anecdotal statements with this sort of hyperbole. Just to illustrate the larger point, I find something wrong with nearly everything Haiku writes, but then again, I don’t expect perfection. I’d probably get a “better” end result for most individual runs with the more expensive models, but at vastly higher cost that doesn’t justify the difference.
> I don't think it's really helpful to tell people they're holding it wrong
I’m not saying that. If anything, it really doesn’t matter much what model you use, and it’s only a case of “you’re holding it wrong” in the sense that you have to use your brain to write code, and that if you outsource your thinking to a machine, that’s the fundamental mistake.
In other words, it’s a tool, not a magic wand. So yeah, you do have to understand how to use it, but in a fairly deterministic way, not in a mysterious woo-woo way.
It’s not snarky. It’s literally the argument people are making: I am special, my use case is exceptional, therefore I need to use the special tool, even if you don’t need to.
Also, GitHub Copilot dropped down to defaulting to GPT5.3 for most requests after culling Opus, and honestly, that was a SOTA model only weeks ago. It blew people's minds about how good it was. And now we're all here moaning. It's even worse than your Ferrari->Bicycle example. It's like they took away your 2026 Ferrari and gave "just" a 2025 Ferrari.
>> Yeah, I hear that a lot, but it never comes with proof. Everyone is special.
You were the one who made the claim that Haiku is fine most of the time. To any reasonable person, the burden of proof is on you. Maybe you should share some high level details about your codebase, like its stack, size, problem domain, and so on? Maybe they are so generic that Haiku indeed does fine for you.
Most of the people using these models aren't skilled enough to make that determination. Seems rough trying to sell yourself as the thing that means you don't need to understand what you're doing but also insist that you understand what you're doing well enough to select an appropriate model.
reply