potamic

I'm not so much surprised about the volatility as I am about all fund houses arriving at the same variation. I would've expected different models to arrive at different predictions and thus smooth out potential big jumps. So I'm finding it hard to understand how every model predicted 25% down instantaneously and simulatneously.

I don't know much about forecasting models and all, are you saying opaque forecasting is some kinda standardized technique and anybody doing opaque forecasting with the same data will likely end up with similar predictions?

Do they all use the same models? How is it that they all arrive on such a large discrepancy with one small, new data point?

BSL knows there two trains are in the block, two have entered and none have left. When LZB is running, it overrides BSL and allows both the trains in the block. But with LZB down and BSL taking over, it should have raised alarms everywhere that there two trains in a block.

afaik, Shinkansen does strict hardware enforced 30km/hr for manual mode. It cannot be overridden.

Yeah, that's what I meant was a failure in the system. Ideally, manual operation or drive-by-sight should be the default mode. Only the appropriate signal state should be allowed to override this and clear the train for higher speed, otherwise the system must enforce manual operation (hardware enforced speed limits etc). There should simply be no procedure for a human to be able to just override this.

Shinkansen does that afaik. 30km/hr cap during manual operation. No exceptions.

During normal operations you can act on the distant signal, because you know you're the only one in the block. But when falling back from linear to point signalling, you cannot make that assumption anymore. You need to get back to a state where you are certain you're the only one in the block, and the only way to do that is to reach the next main signal by sight after which you know you're clear for the that block.

Sure, another way to be certain is for the controller to manually verify that the block is clear and verbally communicate to the driver to override the protocol. But that is putting a lot of burden on human judgement, and even without this incident, anyone would agree is not the best idea.

I don't think this should be required for "normal" operations of the network. Sure, human override should always be there for when an exceptional situation occurs. But you should take extra precaution when that happens because you're already on your final safety net and relying on the judgement of a single human at that time, and that's okay because it's an exceptional situation that doesn't happen often and when it does you can afford to spend an hour or two to take that extra precaution.

The problem happens when the system fails so often that human override becomes the new normal, and because you're doing it every day you now cannot afford to spend the extra hour and tend to skip the extra precaution because you're doing it everyday and start to feel you know what you're doing.

> the signal shows whether you are allowed to drive in the next block, not into the current, as you are already in the current block. This seemed to be what happened here, the second train saw the green signal thought they are back in the automatic safety system and accelerated to 200km/h, not knowing that there is still a train in front.

Don't think the second train saw a green signal. They were in the middle of the block, while the first train was waiting just before the exit of the block (with a green signal ahead).

What I'm gathering from other sources is that the Spanish system allows the controller to override the point-based rules and provide manual clearance (because the blocks are long and it takes forever to reach the next signal by sight). It's a failure of the system imo. Ideally, the system, under no conditions whatsoever, should have even allowed such a practice to be allowed.

Breaking news: Guy on internet reduces an enlightening and constructive article to a triviality.

I'm sorry, but you did it first. I mean, being able to show correlation between variables in a formal, peer-reviewed study is something? What else do you expect scientists to do? This is how science works. You take an assumption, gather data supporting the assumption and establish facts from the assumption. You can't expect researchers to go by anecdotes and vibes when coming up with treatments, can you?

LPCAMM2 sockets were supposed to provide all the benefits of soldering while still being a socket. Anyone know the state of adoption?

I don't think so, I think the roots are deeply cultural. India is a collectivist society where family hierarchy, community loyalty feature strongly, with sayings like "Do not question elders" widely expressed. A lot of this reflects onto work culture as well, among other places, and you see similar things play out across other collectivist cultures as well. So, while arguing with colleagues who are your peers is fine, you will find people will shut down to a large extent when there's hierarchy involved.

To see first hand how this "Do not question authority" plays out, see this every day meeting [1] at one of the largest banks in India. Disclaimer, this video is not AI generated. Sadly it is not even satire.

[1] https://www.youtube.com/watch?v=_SL8-kW7-5A

How big are those verticals? When it comes to management consulting, they enjoy the Big 3 fame, but you don't hear much of them outside of that.

It means they agree they are culpable in the specified activity, and are ready to accept its consequences. The consequences depend on the nature of the activity and the actors involved. If the activity was illegal, they agree to be charged with the crime. If the activity was unethical, they agree to be censured by their peers. If the activity was uncompassionate or immoral, they agree that their standing in community may be affected accordingly.

But most importantly, by assigning blame, they have cleared the contention at hand, taking the critical first step towards resolution of the matter. Without this, the issue may simply be stuck in limbo, taking up valuable time and resources and may never be resolved even. By acknowledging their mistake, they also wish to signal to others that they have learnt a peronsal lesson and would try to avoid making the same mistake in future. So it carries a lot of implications and is not an empty phrase at all.

Sure there may be people who throw around the phrase lightly, whether one believes them or not depends on how they are viewed in the eyes of others. So unless there is reason to believe otherwise, I would give them the benefit of doubt. If you want to imagine a world where the concept of "taking responsibility" doesn't exist, look at politicians. Is that how we want our society to function?

Very interesting!

Could you elaborate on the nature of the work? I'm really finding it hard to imagine what kind of work would require one to work in the dark at all times.

You can move, but if your followers stay on the same AppView which decided to penalize you, you're effectively ex-communicated.

Why does this need a black room with no windows?

What the article above posits, that dark matter could be stripped out from a galaxy after a collision.

The elephant in the room is that the concept of separation of powers has been shown to be a farce. Checks and balances can only work when there is independence between the branches. When a single ruling party is able to influence all three branches, there is concentration of power. And we know very well from history, power corrupts.

This is a problem not just in America but every democracy in the world. Some countries are just fortunate to be blessed with abundance that there's plenty left for everyone even after skimming off the top. But it's a ticking time bomb nonetheless ready to set off when the circumstances align.

I think people need to think about the fundamental structure of government. The three branch system as it functions today in many countries is clearly compromisable. The system needs to change, there must be a better mechanism to guarantee independence and integrity within the branches without the honor system.

> If two typical galaxies met in a head-on collision, the dark matter would experience very little friction, or drag, and would keep on rocketing by. But ordinary matter, which interacts using all four fundamental forces, would get all tangled up.

Isn't a galaxy mostly empty space? I thought I read before that when galaxies collide there's actually very little physical contact between the matter. Stars mostly pass through affected predominantly only by change in surrounding gravity. Is this not the case?

What gets lost more is that Balwani was convicted of defrauding patients, so it's not even like Theranos was absolved of those charges. The kicker is that the CEO and co-founder of a company that has been established to willfully defraud patients somehow managed to escape conviction.

It's perhaps a good time to bring attention to this open issue [1] on vscode from 2015 with thousands of upvotes, hundreds of comments pleading, the collective hive mind losing their minds over how a usability requirement so basic can be left neglected for years altogether even as so many other not so requested features seem to have no problem getting piled on.

While this issue has been a personal pain point, I do recall there were quite a few widely requested usability issues that went unaddressed over the years. So I'm not sure about how to react when I read this blog post about "modernizing" vscode UI in 2026. How can you talk about modernizing and journey through a 5-stage evolution when fundamental aspects are so obviously unaddressed? What is a modern UI anyways and how is it better than ancient UI? What specific advancements in HCI research from the last 10 years are being incorporated here to call the UI modernized?

I have to take a dig at the larger design community because I think this is not an isolated problem. Software usability in general has been going down for me over the years. There was a time when you would hear more from the community about the science of interaction and why you would design an interface in a certain way. Of late I feel the emphasis shifting more towards visual design with things like colors, spacing, border radii etc. with this blog post being a case in point.

If HN were to be launched today, I'm sure it would look nothing like what it does and instead have a more "modern" UI. But, would it be an improvement?

[1] https://github.com/microsoft/vscode/issues/519

If you cannot tell the difference, you cannot tell the difference. Unfortunate as it may be, the standard of proof for criminal conviction must be high. There's a reason that various legal systems across the world have converged on such principles as proof beyond reasonable doubt, in dubio pro reo, intime conviction etc. basically stating that the legal burden of proof is on the prosecution to present compelling evidence of crime and that the benefit of doubt must go to the accused. It is considered so fundamental and so important that it has made it into the International Covenant on Civil and Political Rights which a majority of countries in the world (including Denmark) have ratified. There must no doubt have been painful lessons in history which led humanity to eventually adopt such ideas. We must be careful to forget those and inadvertently take steps going back towards that direction.

This is why you are not an investor. (Cue McMahon meme), A marketing campaign to get eye balls across all the major tech outlets and take up headspace in everyone's trending feed? Oh yes. A potential, closed source alternative to github you can slowly wrangle your users to move to? Hell yeah. An always online centralized design requiring users to stay connected with the server for every operation? Let's go!

This is an ad for her upcoming, self-starred documentary. So nothing much has changed?

mise is great with running multiple versions of a package, something I've found not very seamless with brew.

Allowing dynamic prices only hurts people with less means who will find themselves wasting a lot of time hunting for the lowest possible price. And businesses may claim to charge based on wealth, but what they will do is charge based on desperation. There are also repercussions on wage, like what's the point of a raise, if I'm going to be paying more for expenses and make the same savings at the end of the day? Fixed prices simply eliminate a lot of inefficiency and headache from the system and prevent everyday shopping from being a cortisol ridden experience for the average person.

Besides, what you're proposing does not result in draining of wealth. It simply moves wealth within the capitalist class and keeps it there. If Bill Gates is paying 150k for a banana at Whole Foods, he's just paying himself because he likely owns a big chunk of equity in Amazon. To fix wealth inequality you have to transfer wealth to the working class. It has to be a tax so that the money moves to the public and is spent for the benefit of the public.

Countries like India and Brazil have shown that you can successfully run payments infrastructure nationally. Turns out it's not all that difficult or expensive or at all. India's NPCI processes 500m transactions daily with an annual budget of about 50m dollars. They make a profit over that while keeping the system free for a majority of users.

[1] https://www.indiatoday.in/business/story/npci-rs-500-crore-p...

What's the significance of warm speedup in the benchmarks? For most new installs, cache would mostly be empty, shouldn't the cold numbers be more significant for day to day use?

> If some vibe-coded clone can compete, can manage updates, win market share, produce value superseding the original does that version not deserve it's share?

I don't mean to judge whether the outcome is better for society over all, but just wondering on the kind of impact it can have on the production side for now. Whether people are going to hesitate going open source and guard their IP in more ways than previously.