- New model released, this changes everything (plus thumb scaled metrics to prove it)

- LLMs kinda suck

- The bubble will burst any day now

- Are Plinko machines conscious?

- LLMs kinda rule

- Tech companies doing unethical things

- Someone discovers that, surprise, agents work better when given context (skills! mcp! monorepo! etc)

- Slop that successfully conned its audience and got past the filters

- And so on

This particular article is slop that got past the filters. See the title, see the jargon, see the insufferableness, see the total lack of narrative structure.

I guess that means we're due for a new model pretty soon.

If the author really did one-shot the PR, then the author must have also put in the extremely detailed and tedious work which would be required to tame LLM output for this specific repo, in which case, they actually put in real effort, applyed expertise and thoughtfulness, which is the whole issue at hand.

That said, I don't know if it matters at this point.... On top of Brandolini's law (mentioned elsewhere), what's more likely to happen is that developers' intellect, made available in open source, is just going to be harvested as context for different projects (I'm not taking about training data, I'm taking about devs literally cloning repos into an agent to harvest intellect).

I see two issues with this however, which probably need to be addressed somehow:

- 1 is that, similar to tech, the positions may be seen as cush/low-effort... Like they did half the real work and are going to phone in the rest with LLMs. (Maybe this is already an existential problem, idk.)

- 2 is that domain knowledge is still difficult to aquire... (which may affect whether this manifests, not necessarily that older engineers like myself can jump in)

Nevertheless, if the pattern of dual expertise were to become common, it could open up a lot of opportunities, and is a positive vision of the future, which I think we pretty desperately need.

I think there's probably low-hanging fruit to outsource reasoning from rote read/writes.... Just speculation though, I'm not a token optimization expert.

I would have been very impressed if it made pixel art for cities on demand. If I were able to put in my hometown and see a delightful pixel art memory, I'd swoon and point to it and say "wow wee LLMs are pretty dope aren't they", but instead I'm just thinking this is yet another shallow project which will be abandoned tomorrow.

SNL has never been funny, at least not in my adulthood. For instance, I got chummed by YouTube, and hooked into watching Tiny Horse, which was embarrassing for both the actors and viewers alike. The the skits play like pre-recorded improv.

These match my own thoughts exactly. As much as I don't love them, I agree with the speaker.

For Factorio players: we used to live in Nauvis. It was beautiful but sparse. Then we outgrew it and unwittingly found ourselves on Fulgora. It is abundant but trashy.

And just like in a game, the rules changed and we adapt. That's what makes the game fun. Granted, that it's all our livelihoods on the line (very high stakes), but all things being equal, we have new challenges in front of us that we can either bemoan or accept.

Seems to undercut some of its own claims, the idea that we need to design interfaces for end users is already antiquated.

Ironically, what I'm working on a post-processing hook for colorizing and summarizing responses without degrading the session quality. So "truncating" = summarizing, "quantizer" = char limits and thresholds, "collides" = conflicts, "two oranges" is referring to the "alert level colors" where a second model (Haiku/Sonnet) colorizes text based on the perceived (or suggested) priority of a response's statements (e.g. "just so you're aware, I didn't commit" is fucking useless and it needs to be blacked out).

So the original insufferable statement translates to something like "The code that checks whether a text fragment is too verbose was conflicting with the part that colorizes the text."

P.S. Let me know if there's something out there that exists like this- something that adds a dimension like color or priority-assessments on a per-response basis. So far all I've seen is 2 dozen ~100k starred GitHub plugins that add zero value or make things worse.

Personally I think a web-of-trust Keybase-style solution could work... But buy-in is difficult... you'd need some sort of "seed" strategy to make the app useful alack of 100 (a huge number) trusted friends.

I'd like to give the author the benefit of the doubt, but since we've set all ourselves on being cynical....