Clojurians Log v2

Axios is used in a lot of projects

Man, what is up with all those NPM vulnerabilities?

crazy, right?

one would think they don't know what's going on in their dependencies thinking-face

looks like this is going a daily/weekly occurence for a while IMHO.

I was doing a Seinfeld bit

aaaah missed that cultural reference.... 🙂

never really watched it.

I love Seinfeld

Neither have I!

I love NPM less than Seinfeld, and I've not seen much Seinfeld, at all

The only thing I watched of Seinfeld is Elaine's iconic dance!

I just know the memes

aah the important bits

It happens so often that I no longer tease my friend Laurie about npm security stuff 😉

You get what you f*cking deserve

now consider what kind of technology LLMs will choose for you! Independent, good stuff? Or pre-packaged all you can eat AI buffets? Just wondering!

Your aligned AI, sir

very aligned with my creator's business interests, am I right

Monkey's paw curls

Current bedtime stories in the horror genre.

@ben.sless it was his colleague that invited in the vampire. There are many naive fools in this house. Let's not blame each other.

Let’s have a vampire awareness drive 🧛‍♂️🧛‍♀️

And start nailing the windows shut

I would not have this prejudice against the vitally impaired stand

I'm not averse to LLM usage, but you better know what kind of Faustian pact you're signing

(it won't happen to me (I'm built different))

Wow, that happened faster than I had anticipated. Or maybe it's just a fluke, and the actual enshittification hasn't begun in earnest yet. I don't keep up - I've just read headlines saying that the AI companies are burning money faster than it can be printed, and they aren't estimated to be profitable anytime soon.

I read they'd already backtracked on this due to the online backlash...

They backtracked for now

Heh, fair.

it is very easy to get extracted.... One thing I learned a long time ago was starting at the very first error that occurs... everything after that can be due to the first error.

Indeed! I've been guilty of hunting down consequences of errors instead of errors at times.

And I, too, once only saw word salad where I should have seen a beautifully error-documenting stack trace. Slowing down can help in these regards, I find.

I had this experience very recently! I was diagnosing a problem. A bug, coupled with bad error messages. At first, I quickly wrote some code that solved the bug. But something didn't feel quite right. The place I added my code just got messier. What got me unstuck yesterday was enumeration of criteria and alternatives! In other words, decision matrix. When I took the time to fill known information into the decision matrix, the solution glared back at me. The original wasn't great, and my modified version wasn't great either. I deleted the old validation code (that used to give bad error messages), and worked through validation and reporting for each failure situation (criterion). I feel much better about the design and the code now, and I wouldn't have gotten there without slowing down, taking the time needed, and making sure I really understood the problem.

Well done 🙌

I feel like I should slow down enough to get used to using decision matrices... It feels like an obviously helpful tool.

come to the Clojure meetup in Leuven tomorrow if you want to fight me 😉

Nah, rather not

I'll be there. Do we have to fight? 😅

you can get a hug instead

I'll be there for it 🥰

> We’ll miss the satisfaction of the artist’s signature at the bottom of the oil painting, the GitHub repo saying “I made this.” 💔 No more Hackers and Painters?

PG was always overrated

elaborate?

no

you can tell me over a beer

maybe 😛

nah, I don't want it to be about PG, I did like the hacker/painter analogy though

Good article. I did not know that their ideology had a name (https://en.wikipedia.org/wiki/TESCREAL?ref=gaiwan.co). AI is just another tool and, as one of my colleagues says, I''m not impressed. Its price will go up a lot in the coming weeks/months.

I want to push back on the idea that it's just another tool. That just is doing a lot of work. Compare an IDE to AI. Their externalities are radically different to the point where they cannot imho be waved away.

I agree, I wouldn't call it "just another tool". It is a tool, but one that comes with a lot of questionable baggage (moral/ethical/professional responsability/etc questions). For some (many?) that makes it a non-starter. I guess the point I get to is that, even if you decide to use it, which many people do and which I do understand (I use it too, sparingly, with specific constraints on models/providers), I would still encourage people to engage with those questions and concerns. To inspect their own value system and any tensions within it that may arise. It was somewhat written in a rush so maybe that didn't come across that well.

With great power comes great responsibility... as a wise man once said.

And yes it is a tool, but more like a pair of small scissors compared to a very big and powerful lawnmower... that sometimes does its own thing.

Mowing the lawn is how Israel refers to its regular culls. The problem isn't that machines kill by accident it’s when they kill by policy. And LLMs are tools of fascist regimes today. It needs to be said.

Hadn't realised they use that term.

Sorry it wasn't meant as dig against you. Sorry if that was implied

No, that is fine. I just made up on the spot.

Agreed, degenerative AI is a different kind of tool. What I meant is that it will never become AGI, it's just an useful tool for some tasks. It's shocking to me how many companies have decided to pivot to AI. Is this some kind of collective delusion?

There's a ton of money behind it. Maybe related.