kasperjanssens.bsky.social
87 posts
22 followers
60 following
Discussion Master
comment in response to
post
Oh yes it is. And maybe if I think about it more your first statement about FTX is the right one, that it's more about culture and even in an imperfect company you can do good if your department has the correct culture. Or maybe that's my mid life crisis talking ;-)
comment in response to
post
(for virologists I mean)
comment in response to
post
It's just the annoying part of software, you need to choose where to work and it's sometimes tempting to not care about the implications. But yeah, goes for a lot of fields, biological weapons are also interesting I guess
comment in response to
post
There was a time when the whole world wanted to work in Rust and the only companies who did so and were ready to pay were the proof of stake dudes.
comment in response to
post
Yeah, that's likely true. The sad thing about this and eg Solana is that the technological environments are interesting but at least for Solana for sure you cannot wash the stink of. It's what draws in decent people too. And assholes who don't care as well, of course.
comment in response to
post
I take it this means it's beyond legal ;-)? Could be I honestly don't know, I'll see whether I find some more info about it somewhere.
comment in response to
post
Yeah, that's not great, and their product is, iiuc, on the edge of legal but is it so problematic?
comment in response to
post
It's a really cool technological environment they have, and they pay over the top, but I have to admit I do not know a lot about the implications of their product itself. What is the main issue about it, did you maybe write something about it or?
comment in response to
post
And the money and energy requirements are of course insane, for the large models. But you could imagine a way of having this type of tech without the energy requirements. The VC investments in the LLMs feel more like a ponzi then anything else, so that is totally correct.
comment in response to
post
Yeah, I suspect when the dust settles, the Chinese approach of having a small-ish system that can act as a mediocre junior colleague could have merit. But the large models. In software, one way to measure qualit is if you can quickly fix bugs. In LLMs that is for instance not possible. Red flag.
comment in response to
post
If you dig down, it's also often not true and just a way to make the tech sell. Maybe you knew, wasn't clear which nonsense you were talking about.
comment in response to
post
I agree, but I recall something else you said that struck me, that the Dutch have their back to Europe, facing towards the UK. Could that skew your political identity in a way that it wouldn't in the rest of Europe or do you think it's more common/something else? Same video btw, interesting one ;-)
comment in response to
post
I recall you making this point in a Brexit video, the lack of European Deimos. It's something that I feel is extremely correct, and maybe not even that hard to solve. The EU itself could start a newspaper like medium, not going to be cheap but in the grand scheme of expenses likely not even top ten.
comment in response to
post
Yeah I get it. It's also a fear over here, cause at the very least your whereabouts are recorded every time you pass by even if you did not drive too fast (potentially anyway). You can have a system that 'does not store if not too fast' but if being lazy gives you more surveillance, good motivation.
comment in response to
post
France was even going to introduce average speed monitoring on large parts of their highway, I read last year (holiday prep ;-)). They have great road infrastructure so reasonably easily doable. Scary maybe, but, tbh, a letter with a fine is less threatening than being pulled over.
comment in response to
post
It's not limited to China, fyi. Nearly all of Europe has speed camera's or systems that check whether your average speed over e.g. 10km of highway was too high. On the one hand, it could be abused, true. Otoh, frees up time for police to do policing, and as you say it should be enforced.
comment in response to
post
I do not really see where COTS comes into play in this discussion. The drone, the server, the software, ...? Software-wise, a telegram bot is going to be far more bespoke than any other way of communication. Anyway, it's not so important, just fail to see what Telegram would bring in this use case.
comment in response to
post
Well, that's what I don't see. From my point of view, connecting to that telegram server is not different than connecting to a server that you set up yourself, all IP. But with the one you set up yourself you've got full control over how precisely and who reads. Only disadvantages to Telegram.
comment in response to
post
Sure, but why would you use telegram for this? Technically it's not very practical, and what do you win?
comment in response to
post
Having written a Telegram bot before (granted not one to drive killer drones) I feel that this would be a very clunky way to communicate with a drone tbh. A text based protocol, there is no advantage for machine to machine communication. It all feels a bit off.
comment in response to
post
How would 4G communications be immune to jamming/electronic warfare? Isn't that weird?
comment in response to
post
I am not up to date on which naval assets Frontex currently has but does this not point to more organisations getting involved very rapidly than just Frontex?
comment in response to
post
It's indeed weaseling around 'generated by software' and then saying it's always AI. My IntelliJ generates a lot of boiler plate for me too -> generated by software. Not exactly an LLM though. Thing is, if this is necessary to hype it, doesn't bring a lot of confidence in the capabilities neither
comment in response to
post
At the same time, if I would have 100 euros every time I had to run a bash script that did not even have shellcheck run on it once, I'd be retired by now.
comment in response to
post
Hmm, good question. When I talk to people about scripts, if somebody objects or opposes it, and you start drilling down into arguments, it usually boils down to being less well built, less well tested. Whether this is from out in the world, maybe, maybe not.
comment in response to
post
I'll give it a go: scripting has the connotation of not being tested and if it works, fine, if it doesn't it can end up failing down the line. The point is of course that scripting shouldn't necessarily be without tests, especially not when it starts becoming load bearing. It just too often is.
comment in response to
post
I haven't followed Johan Bruyneel his exploits in a while, but indeed, ever since his suspension he just wants to be in the limelight, as negatively as possible. A few years back in an interview with a Flemish journalist he threatened revealing more Fuentes-names and so on. Never happened of course.
comment in response to
post
Indeed, lots of examples. All those vc backed companies have two possibilities. Or turn into rent seekers asap after completely destroying a market (by operating at loss or break the law). Or sell the company to a larger company as quickly as possible and walk away, bit like crypto.
comment in response to
post
No sorry this is not true. Accessibility is really a completely different topic that should not be folded into a technical course. It's *hard*, it really is. If you expect people in technical courses to have a grasp of this, it's not going to end well. I agree we need it, but it's really separate
comment in response to
post
Belgium has two years of common curriculum for all engineers after which it specialized. I took the comp sci spec but nowhere did it provide a philosophy of sciences course. It's not that uncommon in my experience tbh
comment in response to
post
Well they don't, not always, to be honest. I didn't get one for instance.
comment in response to
post
Well, they don't to be honest. Chemical engineers or material sciences engineers don't. Not always anyway.
comment in response to
post
I fully agree and have been in a similar position. I just feel it's more a failure in the curriculum than a failure to incorporate risks in specific courses. But maybe that's also what you mean. Anyway where it should be tackled might be a rounding error in the whole discussion.
comment in response to
post
I can agree with that, but can we expect people who go deep in a technical topic to teach that? Doesn't it just mean that we lack a science philosophy course in the (packed) curriculum? Cause we seem blame this on a single teacher who maybe should know better
comment in response to
post
But would you then also expect material sciences professors to have foreign policy backgrounds? I agree about the 'being aware of what you're doing' part for the practicioners, but you cannot expect people who teach to be experts in everything that can go wrong with it, or do I misrepresent you?
comment in response to
post
The real drama here is that we need a boatload of infrastructure for the electrification. We do not really need a lot of new tech. But nobody is doing the groundwork of explaining. We will need NPPs, offshore wind, solar where possible, storage,... It's always a social/politics problem in the end.
comment in response to
post
I don't see how that invalidates anything. An extra date representation is fine minutes programming for that same company (putting it between brackets again). Does that mean the rest of the world uses it, no. Not what I said neither.
comment in response to
post
Cfr 'Gulf of America' on Google maps.
comment in response to
post
I honestly think they're capable of renaming 2026 to 1 AT, so it might indeed be the right way to look at it, a very bizarre cult with a very weird God-Emperor.
comment in response to
post
Isn't that the problem with loads of political decisions and directions? It's all fun and games until it needs investment and infrastructure to work. That's why culture war is so tempting, all the goodies with zero infrastructure necessary.
comment in response to
post
I'm not sure they don't know. If you look at the M-series of chips, that's real innovation there, well executed on top of that. I guess they just don't feel the need to innovate in the other parts of their business anymore. True, same outcome, for now, at least.
comment in response to
post
But I agree with the sentiment that not enough has been done by the winners of the last 15 years to keep pushing ahead and delivering more value to customers. Google used to, but not really anymore. The social media barons just started rent seeking asap.
comment in response to
post
I agree with the sentiment about the tech mentioned here being too much hype, but the 'no real innovation' is maybe too much hyperbole. There's a lot of even fundamental theoretical work that underpins all the social media apps that does offer possibilities that did not exist 15 to 20 years ago.
comment in response to
post
One of the problems that at least Europe often struggles with is that it is not always so clear where 'home' exactly is. It often complicates what indeed should be a reasonably straightforward action.
comment in response to
post
In that way. Ok I see. I was a bit confused by your follow up remark about socialism, but it's a term with different meanings. I do wonder how the DEI thing will play out for the US companies and whether it will indeed remain a racket or if somebody will treat it more seriously.
comment in response to
post
I agree, but that goes for nearly any business doesn't it? Most businesses, when they're large enough, like to solve the problem of competition by influencing the laws. Again, far from great, but maybe the least of problematic views in the US currently. But I get your point.
comment in response to
post
Not saying it's great, but isn't this just lobbying, like basically what every interest group does?
comment in response to
post
Leopold II to be precise. And like him, the state will have to fix all the problems he'll leave behind. Likely less long than the Belgian state because less involvement but still
comment in response to
post
Strong 'The Lord of Lies Shall Perish, Chaos Shall Follow in their Footsteps' vibes here.
comment in response to
post
Famously loved the Americans. Often heard talking about du vin, du pain et de l'americain (preparé).