e-topy

I'm curious on your opinion, as a professor, how should universities (not individual professors) adapt to LLMs?

As an undergrad, I hope schools move to educating students to use LLMs in a more responsible way. You can't put the genie back in the bottle, and resisting progress is futile, might as well use the tool we now have to help students learn even faster and better (e.g., making feedback instant and not answers, helping digest or split up material, checking answers).

I know opinions about AI at (not only at) my faculty are very mixed, but I think the answer is going to be in the rational mean, just like how technorealism reacted to the internet[0].

In our last program board sitting, some teachers said that they think programming as a job will be completely irrelevant in two years, while other pushed for more adoption. And meanwhile I know of some students that are basically only passing because of LLMs, and it's bad, like "leaving claude output in markdown files and finished source code on the faculty server in /tmp because opencode did so" bad. And our first year classes completely prohibit even sharing tests or talking about the solutions, which in my opinion a) makes people extremely asocial and atomized b) doesn't prepare students for real life c) promotes dishonesty.

Still, I think our university's thinking is in a stalemate, not wanting pure AI output and useless students, while also wanting to move with the times, and I doubt it's the only one.

[0] https://web.archive.org/web/20081009111415/https://artefaktu... (absolutely amazing read, recommend it)

Maybe the best anti-scraper/LLM protection is going IPv6 only. I'd do that on my website, but I'm afraid some clients might not connect.

Oh yeah I've made some designs for a custom Eurorack style synth, but I quickly realized how expensive it would be to make a full rack. A synth scales exponentially, to get even some usefulness requires getting a lot of modules. Maybe one day.

hell yeah, I have tailed my T-shirts before but nothing else, making my own sound like a nice challenge.

How do you source materials? I'm usually very picky about the material, especially if it touches my skin, and usually the heavier the better. The best T-shirts I ever owned was a military surplus made from organic cotton, and was more than twice the weight of my other T-shirts, but I couldn't find anything like it anywhere.

Oh wow that's getting really interesting, we have a bunch of old and new young apple trees on our field, it'd be really cool making tree mutants.

Do you know any good first resources about this? I'd like to give it a try.

This sounds like a cool mix of genetics and gardening, I like that. Maybe I should ask some local non profits whether they need help with IT infra, because I don't exactly think "google drive and onedrive folders" is going to be an outlier.

also I came back here after three days, how did this get 713 comments all of a sudden???

What you're doing is interesting but those are side-projects. I have plenty of random side-projects, just now after reading Gibson's Burning Chrome, I'm making an OpenBSD server where you can only log in using SSH keys in my implant, and logging in makes you a completely new but very restricted user with 1GB of free storage. Kinda like Johnny Mnemonic.

But I feel very disorganized when most of my attention is on distinct one-off side projects, I want to work on something novel and big. But thanks for your suggestions. It is true that most industries begin when passion oriented people finally meet money oriented people, but most time they are separate.

This reminds me of [0], basically just inviting the most interesting people I know (also transitively the most interesting people they know), and just getting to meet people. I would really like to do this, but half the most interesting people I know are PhD professors I rant with because I'm next to them in a lab. Maybe once my network gets bigger. But I would still like to know more about how you do this, as other people doing this accidentally made me some good friendships, and I'd like to repay this favor to others

[0] https://takes.jamesomalley.co.uk/p/this-might-be-oversharing

I do boxing as a form of cardio so I'm not weightlifting all the time. So I've just invited a friend for a 1v1 and he accepted, time to start training both properly I guess.

I do want something related to computers because that's where I'm skilled the most, but it being mixed with something else is fine (i.e., biohacking). But computers generally are becoming stale, considering how much money has been poured into everything digital, it's going to be hard to find something novel. Maybe the next frontier is becoming an electrician?

No, but they are replacing bad ISPs. I have a relative in Brussels, while there is 10gig fiber on a nearby street, he's stuck on 100/10 coax, and to add insult to injury, Starlink is cheaper.

I don't know how to put it into words, but this aesthetic of computers just looks sooo good. Makes me want to try making my homelab/experimental hardware look like it. Any ideas how to source parts that look similar in low quantities?

You can. GitHub is about to hit zero nines of uptime[0]. But feedback like that is far too late to be useful. Maybe (principal or senior) engineers should be the ones to judge, and be trusted by management that their foresight is worth pushing the deadline?

[0] https://mrshu.github.io/github-statuses/

Git provides file versioning services only, whereas Fossil adds an integrated wiki, ticketing & bug tracking, embedded documentation, technical notes, a web forum, and a chat service [...]

I like the idea of having all of those within the actual VCS, mostly because with Git you need centralized services like GitHub to provide that.

But I have to ask: Is it really a good idea? Seems like feature creep motivated by the wants of a single project (SQLite).

All of those could be (albeit awkwardly) backed with a git repo and a cron job. Wiki? Just make a repo with a bunch of Markdown or this-week's-favorate-markup-language files. Ticketing & bug tracking? Again, just a Markdown file for every ticket. Embedded documentation & technical notes? Those are just special wiki pages with different attributes. Forum and chat service? Do you want your VCS to do that? I get being able to hyperlink between files and conversations, but still.

I really like the idea of making your company's office a publicly accessible coworking space. It does fit Kagi's slogan of 'Humanize the web', imagine if you could go into Google and chat with Larry Page.

This works in RustRover as well! Super useful.

European here. I've been to the US and holy mother of Jesus, you put sooooo much sugar into everything. I had to buy 'European' bread because your normal bread made my gums hurt, and even then that was the sweetest bread I've ever eaten.

Seriously, when your one large oreo shake has 2600 calories, no wonder your obesity rate is 35% and isn't slowing. Driving to the toilet instead of walking also doesn't help. Then your hospitals get overrun with preventable diseases and healthcare gets expensive. This isn't a 'caring' problem when getting fat is the only option for most people, the way most people life is specifically designed to make you obese.

Setting the value of href navigates to the provided URL [0]

It would have been caught because this API (setters) is impossible with Rust. At best, you'd have a .set_href(String).await, which would stop the thread until the location has been updated and therefore the value stabilized. At worst, you'd have a public .href variable, but because the setter pattern is impossible, you know there must be some process checking and scheduling updates.

[0] https://developer.mozilla.org/en-US/docs/Web/API/Location/hr...

Yeah, metric is cool and all, you can divide by ten and multiply by ten. But even better would be a hexadecimal system so that you could halve, third and quarter it. Plus it's n^2 so it's a perfect square \s

Instead of using a new PNG standard, I'd still rather use JPEG XL just because it has progressive decoding. And you know, whilst looking like png, being as small as webp, supporting HDR and animations, and having even faster decoding speed.

https://dennisforbes.ca/articles/jpegxl_just_won_the_image_w...

Jason Booth (the author) also has a YouTube channel talking about similar topics. I really liked his 'Practical Optimizations' video: https://youtu.be/NAVbI1HIzCE

Apple's Face ID uses what is essentially a 3D camera, a simple 2D color camera cannot compare to that in terms of accuracy.

I read somewhere that every augmentation is also an amputation. Progress in tech means we are constantly lobotomising a majority of the population.

Just thought about something:

There are a few sides to this. There is innovation that just makes things easier but doesn't amputate, like typing machines vs word (took me a while to come up with an example, essentially just evolution). Then there are things that are so old it's useless to know them. Like making butter, sure you can do it if you want to, might be fun, but in the grand scheme of things irrelevant. Then there's stuff that is in decline but needed anyway. Like being able to read a book.

Maybe you could express this as a 2D graph, where X is how much people know it and Y is how much people need to know it.

Not in the Czech alphabet: a, á, b, c, č, d, ď, e, é, ě, ...

Also, we regard 'Ch' as its own letter. So yeah, try sorting alphabetically. I'll wait.

And it even remembers what you chose last time and pushes it to the top. That is UX. Being actually helpful and not fucking annoying.

This very much reminds me of Alexander Obenauer's Lab Notes[1] and OLLOS[2]. They're both great, recommend reading them.

[1] https://alexanderobenauer.com/labnotes/000/ [2] https://alexanderobenauer.com/ollos/

This is why I like SourceHut. It's fast, lean, js-free, shows only what's important on screen.

The only downside is that I now need to juggle between two git servers that don't make mirroring easy. But if (when) Microsoft decides to do something stupid, I at least have an escape hatch.

Commenting just because you can't favorite comments

Damn, his other blog posts are stellar as well, nice!

I only wanted to use Smalltalk as an example of a message oriented language. Even then, every IO and every message-based communication is non-blocking and async, therefore programs have much higher throughput (if designed correctly, the same could be said for C++, but the effort to do so is much, much lesser in languages like Elixir/Erlang than C++, and in most situation comes for free).

As I said in my other reply, it was more research focused, and went on to be implemented in.NET.

From what I've read (https://joeduffyblog.com/2015/11/03/blogging-about-midori/), Midori was a research project, and not a replacement. They started with .NET and C# because they own it, but then departed to custom languages. Their research included stuff like language safety (most of it now in Rust, and it seems they implemented some of it into C++ and C#). Also cool concepts like capabilities instead of permissions (2. post), programs that are multiple separate processes (like Smalltalk's; the lessons learned are now in C#'s await keyword), etc. The articles are a nice read.