Clojurians Log v2
clojure 2026-10-03
Nice, I use that a lot too!
you could argue it serves a documentation purpose. but then you could also argue about re-implementing reverse.
for the movie poster, is there a reason they named the lambda? (fn reverse [coll] (reduce conj () coll)) if not that's pretty funny
there must be some reason, https://www.reddit.com/r/Clojure/comments/1smhrax/comment/ogjuv5g/ . I know it's a common example of where static types are not very helpful (ie. string -> string)
Why are there two lambda symbols on belt? Looks AI generated.
two belts and suspenders (very safe)
oof. when you need a better stack trace because you re-implement reverse
ok heading down to pickup at terminal B. but give me a sec i need a smoke
Rich provided the code
named lambdas make stack traces a lot easier to read
@alexmiller out of curiousition, you said not AI -- that was hand drawn by a human artist?
al la "disclose any use of AI"
nice
wow its so cool to see the sketches
now i'm gonna have to talk my wife into watching the movie
the one thing that struck me of this conj was how far apart we are on our perspective on AI. all of the perspectives are valid. I'm a little more "AI forward" -- my job is to imagine architecture, constrain the behavior of the AI (through various means) and read every line of code. just like an engineer on a team
and i write the code that i want the AI to model -- thats an important aspect?
I am not dismissing the art of clojure. Thats still what i love. But if the product doesn't look exactly the way you'd write it yourself, maybe you're not using the AI fully enough
how far apart we are on our perspective on AIIgnoring the question of “is this slop?” do you agree with any of the following • data center energy was growing approximately linearly until about 2024 and then started growing exponentially. • data center energy use is projected to continue to grow exponentially for the next two years or more • exponentially increasing energy use is bad for the environment and communities • spending money on AI is accruing wealth and power to people who shouldn't be given more wealth and power • the enshittification of internet began before AI, but AI has dramatically accelerated enshittification • the work and human creativity that AIs used for training was stolen (in a moral sense and potentially in a legal sense) I think part of the disconnect might be that some folks are only asking “is this slop?” and are ignoring or handwaving many of the likely consequences of long term AI use.
i dunno. thats heavy. but did the invention of the compiler accrue wealth and power? And is it relevant now if you you hand optimize assembler?
i'm not dismissing your opinions for sure. i don't know.
oof i'm going to be more "pro" ai here than i actually am for the sake of argument. but i don't see how this is different. Is it political? google was cool with everyone but AI is not?
answer the questions I posed for google and for AI
they're fair questions, lol. I'd argue the enshittification of the internet started when you couldn't any longer send someone a check after they shipped your item from misc.forsale Usenet group
i don't know how to answer your questions, though. I don't even know how to start the discussion to talk about it. My thing was, as a clojure commutity, we're all over the map and i thought that was interesting from the conj. I don't know how to reduce power usage of data centers
As an industry, most of us probably had zero idea how to make LLMs do anything useful in the beginning. We've since spent a crazy amount of time, money, and energy reducing the slop ratio. I'd feel a lot better if we had spent even a fraction of time considering some of the other questions.
I'm actually not sure how far apart we are. When you mention all these questions besides “is this slop?”, people tend to nod along. It's unclear if there is actual disagreement or if some folks are prioritizing “is this slop?” differently than these other challenges.
there are a lot of ways to constrain the behaviors of LLMs. Uncle bob has one approach. i use a little of that but mostly i've been working on building a memory model. this isn’t polished but here's what i've been playing with
i'm not saying this is the “solution” of anything just saying it’s software i use to develop other software https://github.com/danlentz/sandbar/tree/0.2.0-alpha
i can understand how you can see this from all sides though. i'm not saying you're wrong about anything. i'm just as confused as everyone else
unfortunately, “is this slop?” is such a great nerd snipe that almost every conversation devolves into “is this slop?” while ignoring or hand waving all the other important challenges we face.
so, your issues are more sociological / economic/ecological? i’ve had a few human engineers also write slop, lol
and cultural and political, but yes
I posted the original question more to mean “in the technical sense”. i don't know how to address any of the others
I only meant to explain why there might be a gap in perspectives. If some folks don't find tips and techniques for improving productivity using LLMs persuasive, these other challenges may be why. For me, I don't know how to evaluate software design without considering its effect on the world. I'll agree to disagree rather than further derailing the thread.
not “derailing” at all. these are good conversations everyone needs to start having
Please bear in mind we created #C0B7Q4XRM1U so we don't have AI discussions in the other/main channels.
Anyone wanting to continue this discussion should take it to that channel, or your messages will be deleted.
@seancorfield Threads can't be moved to another channel?
@smith.adriane do you agree that open models like Qwen you can run locally are already becoming very capable and surpassing what frontier models were doing even half a year ago, and that none of the problem you listed apply to them?
@yogthos As per what Sean said, could you please move the question to #C0B7Q4XRM1U?
(that question was deleted—as I warned any further discussion here would be)
Are @dan.lentz and @danlentz the same person?
@grzm yes, he has two accounts
Thanks. That's a bit confusing.