berkes

24 ms·

301.
by berkes 2y ago
There's so much that you cannot learn from the Internet, but must practiced, coached, steered, etc. That needs fysical things to interact with. That need teams, colleagues, or other humans. People who think you can learn "everythi
302.
by berkes 2y ago
> to allow/expect devs to add unit tests For me such gigs are a red flag and immediate turn down (I'm freelancer with enough opportunities, luxury position, I know). I would consider it really weird if management dictates exac
303.
by berkes 2y ago
Two things: This requires a language that a human can give the specs in. A language that is both precise and unambiguous and understandable by humans. A... programming language? And it requires AI to understand the effects of changes it mak
304.
by berkes 2y ago
Another problem with lingering warnings is that it's easy to overlook that one new warning that's actually important amongst floods of older warnings.
305.
by berkes 2y ago
The last point is the key. It then creates immense value by avoiding a lot of risk and uncertainty for little effort. Getting from "thousands of warnings" to zero isn't a good ROI in many cases, certainly not on a shortish te
306.
by berkes 2y ago
AI might be heading towards a situation where you can "tell it to build a full featured browser". But this narrative is naive at best, and frankly, getting annoying too. Especially because it gets repeated by people who know even
307.
by berkes 2y ago
> It will implement a spec that's written in English or another human language. No, it won't. Because "human languages" lack the precision to describe such a spec. This is exactly why programming languages exist in th
308.
by berkes 2y ago
Not just with SQLi, but I've managed to statistically proof "information" with timing attacks. Where if you join another table (by e.g. requesting extra info in a graphql query) the response goes from ms to s or even m. Indic
309.
by berkes 2y ago
> the only solution is what is discussed at the end: updating the law. That, and actually penetrating the data system and subsequently "leaking" parts of it. Which is nearly always illegal, but could be considered a form of &qu
310.
by berkes 2y ago
It's very unrealistic. There aren't any OIDC and other authentication flows that involve at least five redirects and intermediate login screens. Nor are there any T&C and data processing dialogs. I'd expect at least one &
311.
by berkes 2y ago
I think many of the "AI can do coding" narratives don't see what coding means in real situations. It's finding out why "jbdoe1337" added this large if/else around the entire function body back in 2016 - it
312.
by berkes 2y ago
I would imagine that it'd be best done with the "share" API that already exists. A website designer then might say "share this url on click" and the browser takes over and/or defers to the OS. Now, the share()
313.
by berkes 2y ago
> But what are the parts in the digital world? In essence, all software is a service; often legally even. But for this we'd need to look at software that's "shipped" to all customers in more or less the same form. But
314.
by berkes 2y ago
It's not pedantic and actually a good question, IMO. One that, IMNSHO, isn't discussed enough. WASM esp in the browser offers a great opportunity to "Do Things Right" that JS in the browser got all wrong in hindsight. I&
315.
by berkes 2y ago
Hergé took a lot of effort to get details right. Not just the drawings¹, but also the layout and lettering of the speech bubbles. I own all TinTin comics in Dutch (some old collectors items) and a very few in French. Dutch is often a lot lo
316.
by berkes 2y ago
Yes, but it seems > No JavaScript is sent to the browser. Is a design goal. I doubt it is possible without JS. Especially inside SVG.
317.
by berkes 2y ago
You cannot. Because these aren't simple "URL parsers" or such. They commonly use heuristics, bayesian logic and complex statistics (the word AI has become so conflated with LLMs and GPTs, and infected with politics, but it is
318.
by berkes 2y ago
"yet". > And at present, there's really only one each of the Relay and (Twitter-like) AppView components used in practice, but there is a real possibility of this changing and real architectural affordance work to allow i
319.
by berkes 2y ago
Unless they interop. And there is no reason why they shouldn't. (other than grumpy old neckbeards bickering over their Prefered Thing and blocking any progress towards cooperation or interoperability. Which is actually a thing in &quot
320.
by berkes 2y ago
bsky is decentralized, in theory more so than Mastodon, which is federated (a variant of decentralization that requires you to trust the server admins of a server of your choice) "In theory", because so far, there are very few nod
321.
by berkes 2y ago
I'm not defending Twitter or their policy in any way (disclaimer: I left Twitter the moment Elmo took over. I despise his hypocrisy and his fascist ideas) But this could be a "legitimate f-up". Normally, most of these unsafe-
322.
by berkes 2y ago
Oh, mine is certainly not even close to that one :)
323.
by berkes 2y ago
I deliberately did not make the maze in a shape, but kept it square. I understand that everyone's taste is different. But I went for a form that's visually appealing (to me). As mentioned in the about, I was inspired by screenprin
324.
by berkes 2y ago
Ooh. Do you mean that if you solve it, it resembles a letter? I'm not aware of such maze generation algorithms. Any pointers? I guess it's mostly a matter of steering the random wall selection initially to form a shape, then cont
325.
by berkes 2y ago
I wasn't aware of this whole class thing. Is it an American thing? Certainly a fun idea. Though not that trivial as the entire app now runs client side only. I'd have to think this through, as it should be possible.
326.
by berkes 2y ago
You sound like you know it to be trivial to build this. If you can build it, (which I doubt) please contribute to OSM by getting involved with e.g. HOT OSM. https://www.hotosm.org/get-involved And I doubt it is that easy, b
327.
by berkes 2y ago
It'd be even nicer if the maze is potentially (un)solvable in that case: you'll have to keep trying to find out if its a friend-zone.
328.
by berkes 2y ago
I gave it some more thought and want to take it a step further even (but lacked the time to polish before valentine) The one you land, indeed a hash. It is read only, but allows you to trace the maze with the mouse or touch. Once solved, it
329.
by berkes 2y ago
Thanks! I had a version where I experimented with ways to find the furthest item at the border, but it added little and was very convoluted. For one, to solve this exact issue. Then I realized that the "bug" could easily be a &quo
330.
by berkes 2y ago
Interesting idea. I didn't consider that there could be a difference between the link that was shared and the place where you make yours. I just landed on the version that has both in one by happenstance. But the idea makes sense. It w

More ›