alxlaz · HackerTrans

3 miesiące temudiscuss

It's very hard to untangle it from the rest of its context (the prompt is built dynamically, from a lot of parts, some project-specific, some specific to my preferences, some built from interaction history), so I can't really share it. In any case, I don't think it's some specific prompt engineering sorcery I'm doing, it's not like I've spent any real time refining it or experimenting with various magical incantations. It's probably just some model features making it more amenable to the kind of instructions that are relevant in these cases (directness, questioning trade-offs, thoroughness etc.). My chatbot swears equally graphical in review prompts and news summarizing prompts so I'm pretty sure I'm not tickling the machine just right :)

alxlaz

3 miesiące temudiscuss

I use their API for several models, both for personal and professional use. I think their approach (smaller, specialised models that are well-adapted for specific tasks) is a very good fit for how I work. And even the more general-purpose ones, like the chat model, just... refreshingly good in a lot of ways. My "ruthless review" prompt, which I use for, well, ruthless, guided reviews of early technical drafts, has good technical results for early reviews and holy crap is it ruthless and does it know how to swear. By the time Claude or ChatGPT are done being honest about how right I am to push back and gently circling back, Mistral's large model has sent me back to the drawing board twice.

Being in the EU does smooth a lot of things in terms of compliance, payment processing and whatnot, but I also like that their data retention and privacy policies are pretty clearly spelled out. I need to know something, there's a good chance it's explained outright somewhere and I don't need to read between the EULA lines and wonder what it means.

I do hit limits in terms of capabilities sometimes, and I'm sure other providers' services offer better results for some things. But the businesses ran on top of those more capable models feel too much like a scam at this point and I'd rather not depend on them for anything I actually need.

alxlaz

w zeszłym rokudiscuss

If we're talking about the Budapest Memorandum, cirumstances possibly changing are actually part of the treaty. The memorandum conditions the use of force against Ukraine, and makes provisions for what would happen if a party to the treaty or one of its allies attacks another party to the treaty (second and fifth point, one is about conventional response, the other about nuclear response):

> [The parties shall] refrain from the threat or use of force against the territorial integrity or political independence of the signatories to the memorandum, and undertake that none of their weapons will ever be used against these countries, except in cases of self-defense or otherwise in accordance with the Charter of the United Nations.

If memory serves right, Ukraine didn't invade Russia, and Russia is not under a UN mandate.

alxlaz

4 lata temudiscuss

It's worth remembering that much of the churn was caused not by having to rewrite configuration, but by the multitude of bugs that plagued systemd early on, and the way they were treated did not help. systemd upstream was extremely difficult to work with and made a lot of people reluctant to embrace change, because when something didn't work -- which we all understood was natural for software developed in the open -- you were pretty much screwed. Shell scripts and duct tape were bad but it could take less time to fix something than it took to get people "up there" to admit that they're really looking at a bug, not a feature.

I ran systemd very early on because it gave me some things I liked from SMF but I completely understand why so many people hated it. It's a (suite of) program(s) with good ideas that was hampered by extremely poor maintainership and relation with other open source projects.

alxlaz

4 lata temudiscuss

It's a combination of factors but the gist of it is that they're not preventing anyone from doing it -- nobody wants to do it because they can't justify it.

Designing and selling long-lived consumer products is a tremendous effort in terms of regulatory activity and logistics. And if you're just entering the market, you're entering a market where established players have been sitting on standardizing committees, lobbying regulatory bodies, and have had special deals with manufacturers, transporters, TV channels and shopping malls since Jimmy Carter was in the White House, and have been able to leverage them when dealing with online shops since the days of the dotcom boom.

Furthermore, getting that "more of the market share" requires actual first sales to happen in the first place. Selling fewer, better, and inherently more expensive products by claiming that yours are really good and last a long time is not very straightforward at all: it's not like everyone else is openly advertising that they're selling cheap junk that's worse than their 1980s versions in every way except maybe power consumption (because we can make better power supplies now). If you're selling kitchen stoves that last for twenty years, it's gonna be at least seven or eight years before anyone really believes you.

So all this requires tremendous initial investment in a pretty hostile environment. Consumer markets are already perceived as low-yield and under tremendous price pressure. But, worse, some segments -- household appliances in particular -- also suffer from a "marketshare = dead market" problem: if you sold someone a washing machine that lasts fifteen years, you're not going to sell them another one for fifteen years. So if you're selling things that last for fifteen years, you gotta convince the folks whose money you're depending on that you can keep selling things at a reasonable pace for fifteen years from now, and you won't really know if you've succeeded at your basic value proposition ("we make better things that last longer") for at least seven or eight years into it.

That's not just pretty risky, it's on a timeframe that's about an order of magnitude higher than companies/investment funds/people willing to throw that kind of money into things that run on electricity are able to work with.

That's not to say no one has been entering these markets. They have, but they've usually done it through other value propositions -- stuff that's even cheaper, stuff that's ethically sourced and so on. New premium brands do pop up but they either target prosumer/super premium markets (slang for "super rich people who want to brag or impress"), where the rules are slightly different, or go straight for the professional market first, where the rules of the game are very much different, and use whatever reputation they can make there to start selling home appliances too.

alxlaz

4 lata temudiscuss

These ideas aren't born in the actual product design teams, where "don't put touch control near things that get hot" or "no unprotected touch controls on surfaces where spilling can occur" are not concepts that need to be explained. The way this usually happens is that product management hands down the requirement from down top, either on their own or after contracting some expensive industrial design agency (i.e. a mixed team of engineering and art majors who never had to sell, service and especially provide tech support for, any of the products whose outside they designed) to help "refresh" their product line.

Then, when in a shocking twist of events it's absolutely miserable, the tests are slightly mangled to highlight the usability gains ("easier to wipe", "looks modern and integrates well in a modern kitchen") and ensure no one's yearly bonus is endangered.

Surprisingly, it works out really well despite competition because everyone's doing it pretty similarly.

If your next question is "but why on Earth is it done like that!?", well, there are many reasons for that. But the tl;dr is that making kitchen appliances with universal, "boring" features, which last for 20 years, would spell disaster for a lot of companies in a lot of industries, from home appliances to furniture, and for smaller, specialized suppliers in these fields. The market has fine-tuned itself for frequently replaceable junk that's just good enough to that someone can plausibly say "well ackshually" when I refer to it as junk. Anything better would inherently result in much longer growth cycles, which Western management and stakeholders are not capable of handling anymore, and would disrupt "fashionable" trends, which would further prolong growth cycles in both home appliances and connected industries (furniture, home decor, cooking vessels believe it or not, and so on). It would also make it difficult to outsource mass production to cheap factories with poor quality control processes, thus driving prices up even more than the better materials and design would warrant, but that's more of a second-order effect than a cause by now.

Sauce: I used to work on consumer products years ago -- not kitchen appliances specifically but I know people who did those, too. Attending meetings was a lesson in cognitive dissonance and doublespeak and it's one of the least rewarding tech jobs you can imagine.

alxlaz

6 lat temudiscuss

This is pretty much the same in EU, modulo some controversy about the exact point during the course of higher education when you should be expected to be able to design things on your own.

I have no idea how an EE curriculum doesn't cover bypass capacitors. They're in virtually every real-world circuit. Even if it's not an item that's specifically covered in a course, lab or seminar, there's no way you can put a real-world schematic on the projector and not run into one.

I can't point at a specific course I took where they were covered but I am sure everyone who made it to the third year knew what they were. I definitely remember talking about them extensively in at least three courses (Circuit Theory, Digital Circuits and Digital Instrumentation).

alxlaz

6 lat temudiscuss

Maybe this is a problem of semantics, or something that I simply can't quite understand any longer because I am an electrical engineer (or at least I have a diploma that says I am :-) ). But this distinction is very strange, and I'm not sure why/how you're making it. Electronic engineering is nothing like that.

Knowing rules of thumbs and how to "combine components" into something useful is not a parallel skill set to electronic engineering, it's a part of electronic engineering. Rules of thumb, models and reference schematics (from manufacturers or whatever) are pretty much how you come up with the first draft of a circuit.

Also, electronic engineering consists precisely of designing circuits. Quantitative analysis is a big part of it because without it you can't always know whether the circuit you've designed works according to specs (and in practice, yes, experimentation and testing plays a big role, but there are only so many prototypes that you can blow up before you run out of time, and only so many things that you can test for in a regular lab). The whole point of one's activity as an engineer is to combine components into something useful.

Sure, you don't sit in a lab drawing schematics and building things all day, because there are a lot of other activities that go into building a good product. There's a lot of validation work and a lot of analysis and a lot of planning. Some engineers focus their time only on one of these things, especially because they really are so complex, and so complicated, that you can spend a lifetime studying just one of them and you're still left with a lot of stuff to know.

But at the end of the day, designing electronic gizmos is pretty much what you do, regardless of what role you're playing there.

You can certainly make things by sticking components and circuits together without a thorough understanding how they work. But the idea that engineering somehow mostly about understanding a system through mathematical modelling before building it, and design is mostly about making things by sticking components together by intuition, is very much absurd.

alxlaz

7 lat temudiscuss

Not parent but it's something that's been bugging me for a while (especially since I work in a field where you often have to debug based on field logs from three weeks ago from a device halfway across the world). My two cents, I think telemetry is OK if:

* It has a retention policy that ensures a data breach five years down the road won't compromise five years' worth of relevant data. 30 days of full data + maybe an year of relevant analytics (but no detailed data) are more than enough to improve software. Five years' worth of data, even if it's not strictly speaking personal data, can make identity theft trivial.

* The processes and criteria for who gets to access that data and in what manner are clear, transparent, enforceable and can be subjected to appeal by any user. Can I be reasonably sure (i.e. based on public, legally-binding statements from Microsoft) that, for instance, a Microsoft employee who has a beef with me won't be able to stalk me based on that data?

* For a large system or a piece of hardware, if I can have a reasonable assurance that, at least in the short-term (1-2 years), the data won't end up being used for ad targeting. An OS or a piece of hardware is something that you don't switch that often. If I make a purchase, I'd like to be sure that the reasons why I made it remain valid for a while.

"Makes the product better" is a great reason and I'd wholeheartedly support that, as long as I had an assurance that my data is handled responsibly.

I'm just sayin' -- if I break the terms of doing business with Microsoft (e.g. by using a pirated copy), there's a good chance that some DMCA organization comes knocking at my door and the best I can hope for is an out-of-court settlement that ruins me. I sure as hell expect that, if Microsoft breach the terms of doing business, someone can go knocking at their door and the best they can hope for is an out-of-court settlement that ruins them. That's why I insist on "legally binding". A blog post that says we totally don't spy on you is something that you can breach without any real consequences.

Until that's the case (or, you know, until there's no data being siphoned...) I can do my work on Linux just fine, I don't need no damn WSL :-).

alxlaz

7 lat temudiscuss

Yeah, I don't know if I gave the right impression with that last post. I don't think Plasma 5's non-touch experience has been going down disastrously lately. There's nothing that screams "made for mobile" that you can't change (even Breeze's huge widgets, I mean, there's always other themes). And if touch devices are fashionable and is what gets people interested in KDE and gets contributors on board, I'm by all means happy if that gets to be the default :).

But I'm not convinced the "not at the expense of a no-touch interface" part is going to stay true for long.

FWIW:

> Settings Manager: on my machine I can configure it to have the tree view in its hamburger "settings" button, at runtime; no recompile required.

Some distros still enable that feature but it's a distro-specific thing. I don't know if it's maintained anymore.

> Similarly, I don't think you can chalk the change in Virtual Desktop interface up to optimizing for mobile at the expense of desktop. You still have a spinner for "rows", so it's not like it's an effort to get rid of spinners.

No, but it's kind of awkward to use :). Clicking "Add" four times gives you four desktops with the same name, for example. It's definitely not an easier interaction model than the spinner-based one.

...but this sort of discussion (is it better to have an extra spinner, or an extra button and manually edit each desktop's name? Which one is more intuitive? Which one is more discoverable? Which one is etc. etc.) is kind of a bikeshedding dead end to me. As long as it receives bugfixes, as opposed to rewrites, for the foreseeable future, I couldn't be happier.

The bit about bugfixes vs. rewrites isn't just whining, it's kind of a pain point for small-time contributors -- which, realistically speaking, is how 90% of independent developers get into a project, we're all small-time contributors first (unless you're hired by a big company to work right on something full-time, which is increasingly common in the FOSS world, but not specifically for desktops). It's pretty hard to motivate yourself to contribute a fix when you know it's gonna be useless one or two years from now. Feeling like you're participating in the steady improvement of a thing is pretty nice. Feeling like you're participating in the perpetual churn of an eternal beta isn't much fun.

alxlaz

7 lat temudiscuss

They're a world of difference apart! TDE uses its own fork of Qt 3. Qt 3 is huge, and while Timothy Pearson, who maintains it, is an extraordinarily capable programmer, I doubt he and his team can maintain it and TDE that well. LXQT, on the other hand, can be compiled against the latest Qt 5 libraries. Besides, TDE isn't "just" the desktop, there's a whole application suite in there, too. I doubt that you'll get proper, 2019-level support for TLS, for example, in those applications. Bugfixes are occasionally committed but whether or not they're enough is anyone's guess.

Plus, you get all the usual problems, like inconsistent theming (Qt 3 engines, unsurprisingly, don't work with Qt 5; Plastik, CDE, Motif and Windows are in both Qt 3 and Qt 5 but you need to manually add color palettes etc.)

Frankly, even though TDE pushes all the right nostalgia buttons and even though I instantly feel better about anything with Pearson's name on it, I don't think I want it on my systems :). I tried it and it's fun but a bit difficult to use in 2019.

They're in the same space in terms of requirements but a very different space in terms of bugfixes, compatibility and perhaps security.

alxlaz

7 lat temudiscuss

I haven't used lxde so I can't tell you much about it. What I can tell you is that it has a "start" menu, a taskbar, a desktop switcher, icons on the desktop and Openbox :). It doesn't do anything that you haven't seen before, but pretty much everything it does, it does reliably.

It's... I dunno, it's like FVWM95, only from this century. I used Openbox + a bunch of tools cobbled together before. It doesn't do anything my old setup didn't do, but it sure is more comfortable. No scripts, no custom setups... all that was fun twenty years ago but I'm not a teenage l33t h4x0r anymore, I got work to do nowadays...

I'm not sure why people conflate it with Lubuntu, you can use it on any distribution, and the fact that it's very easy to package means there are few packaging-related problems with it. It's also pretty easy to compile from source if you need that, for whatever reason.

alxlaz

7 lat temudiscuss

> Can you cite an example of this?

Sure, several :). For example:

- The tree layout of Settings Manager is no longer around (it can still be enabled at compile time, though, and I think e.g. SuSE still does it). The default one is the strange hierarchy/screen-based which is pretty obviously meant for touch interfaces.

- The same layout is used by Discover and it's very obviously a mobile design. Actually, anything Kirigami-based is like that :).

- The new Virtual Desktops settings page in System Settings is very obviously reworked from the same perspective. Instead of two spinboxes (number of rows, number of desktops), you now have to click "Add" to add a new desktop, then manually edit its name (by default, they're all called "New desktop").

- The default Alt-Tab switcher, which is remarkably awkward to use from the keyboard (switches applications and brings windows to the front at every step, somewhat like Fluxbox' alt-tab if anyone remembers that) is actually very smooth to use on a touch device. You can sort of see that on a desktop, too, if you try to use your mouse the way you'd use your finger on a touchscreen. You can hold-to-scroll through the left-hand side view and switch to a given window by clicking its thumbnail.

(Edit: plus the usual suspects: oversized widgets and titlebars, large icons with humongous space between them etc.)

The good thing is that Plasma is super configurable. I don't mind changing default settings that I don't like. In fact, I'm all for fashionable defaults, I completely understand the dynamics involve there.

But I also don't trust that I'm going to be able to change these default settings 12-18 months from now -- not to settings that are appropriate for a desktop machine with a large monitor, in any case. And KDE is a big beast. It's hard to migrate settings even between two computers running the same KDE version. If you go all in, it's hard to turn back. I've already done that once with KDE 3.2 and it took me months to sort it out when 4.x hit the market. I'm not sure I want to do it again.

Edit: FWIW, I do try to keep an eye on it because it's actually the only Qt-based desktop environment that's unlikely to get abandoned soon :). Plus, while Plasma 5 feels a bit bumpy to me, it's definitely better than KDE 4, and that's a big deal. Last night, in fact, I tried my hand at hacking on Breeze a little, to make it slightly more compact. It's definitely better than other flat themes but boy is it awkward to use with those huge widgets. If I can get it to look okay, I'll try to see if there's a way I can get this into upstream, too (maybe make some things configurable?), or just package it separately as a compact theme for me and anyone who's interested.

Historically, the KDE community has been super friendly and willing to help newcomers. I'm not sure if it's the same now that there's a visual design group and whatnot, but I'd definitely rather write code than whine about software I get for free :).

alxlaz

7 lat temudiscuss

Ah, no, I'm using lxqt at the moment. It's not perfect but it's okay, and I trust its developers' common sense. Every once in a while I fire up Plasma 5, too, but I'm not convinced. In the last year or so it seems to have gone down the mobile-inspired UI rabbit hole and it's getting increasingly awkward to use.

Other than that, I use the same kind of applications you'd expect any desktop user to use, from file managers to text editors and from office suites to web browsers. Only most of them are Qt-based :). The only GTK applications I still use are Firefox (GTK-ish) and Emacs, which can still be compiled against GTK2 which is a good enough compromise for me.

It's nothing personal, I don't want to get into a big rant about how Linux used to be about choice and about how Gnome is a Red Hat conspiracy, or whatever flag is being waved in 4chan & friends. I just don't like GTK3's UX choices. Oversized widgets, awkward scrollbars and hamburger menus aren't pleasant to use on a 30" monitor. A few years ago I tried to get around that with a custom theme, but theming is... kind of frowned upon in GTK land, so it didn't end well.

But, you know, their code -- their choice. I don't like the choices but I think it's a big step forward from the '90s and a big test of maturity for the open source community. Making (what I believe to be) our own mistakes is way better than cargo culting Microsoft and Apple, which is what we've done up until 5-6 years ago.

Edit: also, why are you folks downvoting the parent comment? It's a productive question, even if it looks uninformed (i.e. I'm guessing the downvotes are because it's conflating "GTK applications" with "desktop applications", which looks silly but have you had a look at Ubuntu lately?). If you think that's a stupid question, fine, but are you really that sure you've never asked a stupid question in your life?

alxlaz

7 lat temudiscuss

I stopped using GTK applications a while ago, so I don't use XFCE anymore, but I still love it, just... from a distance. It's an increasingly uncommon breed of software, which gets better with every release. No pointless shuffling of user interface features in the name of UX "innovation", no rationalizing bugs so that they look like deliberate design choices, just steady improvement. All this by an independent team with basically zero funding.

IMHO, XFCE is one of the few relatively well-known projects that still embodies the spirit of free software as we once knew it. Kudos to the quiet and tenacious people behind it, who still manage to come up with one quality release after another, so many years down the road.

If you want to support (at least some of) their work, Sean Davis has a Patreon account: https://www.patreon.com/bluesabre . Sean is one of XFCE's core developers and the Xubuntu tech lead. I don't know if there's a more official XFCE donation sink (this one is the only one I know of), but hey, it would be a great start!

alxlaz

7 lat temudiscuss

> Be it ordering some product online or booking a flight or other travel ticket or ordering a service

That's true, but all these places already track the living hell out of you. Even the newsletters they send over email have tracking information. By the time they've sent you an email after your first purchase, they know everything they need to show you relevant ads (in fact, that's probably why you made the first purchase...). I doubt bulk analysis of emails can show anything that is not already known way before the emails got sent.

alxlaz

7 lat temudiscuss

It's just the passage of time that has rendered it misinformation. Until not so long ago, Gmail messages were actually scanned for ads -- IIRC, Google was actually pretty upfront about it when they first launched Gmail, and explained that it's how they could afford to give users 1G of inbox space in a day and age where 25 MB was pretty good and 100 MB was pretty hard to get for free.

They eventually stopped, although the phrasing of the privacy policy is vague enough that, as wodenokoto mentions above, I wouldn't be surprised if email messages were still scanned for some advertising purposes. The fragment on the page you link to is only about ads shown in Gmail, doesn't exclude using keywords and messages for tracking, classifying etc. (it just doesn't use them "to show you ads") and doesn't actually exclude using programs to process messages (i.e. you can still reasonably say that "nobody reads" messages if you just feed them into a program). It's also not very obvious if "messages in your inbox" also includes messages you send.

FWIW, I think the policy is deliberately open-ended so as to be future-proof, but I doubt emails are an important source of advertising data today, so I think it's likely that Google doesn't rely on it that much anymore. Most sources of legitimate (i.e. non-spam email) that are useful for advertising -- e.g. online shops and the like -- already track you and show you ads, and Google is already deeply embedded there. Millions and millions of personal accounts are an useful strategic asset to have but I think there are better sources of data.

comments