MattyRad

14 時間前議論

This is an agreeable article, sure, but the idea that an LLM fueled group or team will collectively have this discipline is... idk... bemusing and saddening at the same time.

MattyRad

19 日前議論

"Involution" is spot on.

MattyRad

28 日前議論

Not to distract from the message, but I appreciate that this is largely plaintext not React vibeslop.

MattyRad

28 日前議論

My first thought is that this government-Anthropic feud is good publicity for both of them.

  - Anthropic is seen as a victim/hero
  - They get Government-endorsed model hype
  - Monday will be a bad publicity day with the new Agent SDK limits, this overrides/dominates the headlines

  - The government gets to appear like they're ahead of the curve
  - The government gets to appear forcible and weapons-conscious (and maybe earn some right-wing points)

The government is possibly a real threat here, but it's also possible that this is a case of knights rallying the mooks (https://ribbonfarm.com/2020/01/16/the-internet-of-beefs/), and the models will be back online Monday with a note that "we gave em hell in court because we're so smart and dedicated and talented and good at beefing"

MattyRad

先月議論

Also got this email.

I doubt that Anthropic themselves are pushing identity verification, for what that's worth (which is admittedly not worth very much).

It's a real shame though. Time to shop around, or bite the bullet and buy a real GPU.

MattyRad

先月議論

Keybase. Keybase was/is a trust network that was unceremoniously and unduly strangled. (You were the chosen one!)

I don't know how its style of trust systems can help us solve these major trust problems, but I feel that it's the right direction to save us from the onslaught. If I had the time, this is where I'd focus my efforts, i.e. creating a (maintained) trust overlay on existing social networks. Using slop vacates trust, share your trust signals with other people you trust.

MattyRad

2 か月前議論

Welcome to the dark forest.

MattyRad

2 か月前議論

I think the author is describing the new incarnation of the Death March. In the Death March, contributors know that an active project will be dead-on-arrival, or cannot be redeemed. Maybe a small difference here being that the AI-equipped contributors won't be aware of the project status (i.e. futile).

Maybe this means AI has democratized Death Marches.

MattyRad

2 か月前議論

Good for the author for finding some success. I'd recommend seeking a significant other, somehow that sounds less daunting that making friends past 30. Cool roommates are friend-ajacent and help with loneliness; I had a cool roommate for a while until he moved in with his girlfriend, after which I was deeply lonely until I met my now wife.

MattyRad

2 か月前議論

See below that:

'Don't believe self-serving lies about technologies being "inevitable" or "here to stay". You don't have to just go along with the dominant narrative. You can make deliberate choices and help others to do the same.'

MattyRad

4 か月前議論

> One engineer captured this shift perfectly in a widely shared essay, describing how AI transformed the engineering role from builder to reviewer.

I stopped here. Was this written by an an LLM? This sentence in particular reads exactly like the author supplied said essay as context and this sentence is the LLM's summarization of it. Nowhere is the original article linked, either, further decreasing trust. Moreover, there's an ad at the bottom for some BS "talent" platform to hire the author. This article is probably an LLM generated ad.

My trust is vacated.

This makes me feel that the SWE work/identity crisis is less important than the digital trust crisis.

MattyRad

4 か月前議論

Hoenikker had been experimenting with melting and re-freezing ice-nine in the kitchen of his Cape Cod home.

Beautiful, perhaps like ice-nine is beautiful.

MattyRad

4 か月前議論

Yeah, agreed. I probably wasn't going to delete my OpenAI account (ala the link that is also being upvoted on HN), it just seemed like a hassle vs ceasing to use OpenAI. But when the staff at OpenAI employ mental gymnastics, selective hearing, willful ignorance, or plain ignorance to justify compliance with manmade horrors, I think it's probably important to vote with our feet.

MattyRad

6 か月前議論

I use agentic LLMs (for side projects, properly sandboxed) as much as the next guy, but collectively the normalization of deviance is pretty apparent and shocking (https://en.wikipedia.org/wiki/Normalization_of_deviance).

Backdooring you own machine, sending your .env files, normalizing slop in code review, leaking IP (which is trained on and send to RLHF), YOLO mode, these things would have been unconscionable 2 years ago.

MattyRad

6 か月前議論

Tangentially related, "slop" really isn't a negative enough term for unwanted LLM garbage. "Slop" which is fed to pigs, has utility. "Slop" as a verb doesn't necessarily have a (strong) negative association ("It was slopped on the plate, but it was tasty").

I use the term "barf" more often. Barf has no utility*. Barf is always seen in a negative context. Barf is forcibly ejected from an unwilling participant (the LLM), and barf's foulness is coerced upon everyone that witnesses it. I think it's a better metaphor.

I know that this is just semantics, but still.

* even though LLM output __can__, and often does, have utility, we are specifically referring to unwanted LLM output that does not have utility. I'm not trying to argue that LLMs are objectively useless here, only that they are sometimes misused to the users' detriment.

MattyRad

8 か月前議論

I think this was a really good article up until `on power`, where it became whiny, vindictive, and aimless.

MattyRad

8 か月前議論

Ok, I'll bite though, let's try it out as a non-maintainer.

I loaded https://0github.com/laravel/framework/pull/57499. Completely random, it's a PR in the last github repo I had open.

At 60%, it highlights significantly more test code than the material changes that need review. Strike one.

At no threshold (0-100) does it highlight the deleted code in UniqueBroadcastEvent.php, which seems highly important to review. The maintainer even comments about the removal in the actual PR! Strike two.

The only line that gets highlighted at > 50% in the material code diffs is one that hasn't changed. Strike three.

So, honest attempt, but it didn't work out for me.

MattyRad

8 か月前議論

Seems like a catch-22. For codebases that I'm highly familiar with and regularly perform code review in, I'd say "thanks LLM, but I don't trust you, I'm more familiar with this codebase than you, and I don't need your help." For codebases that I'm not familiar with, I'm not really performing code review (at least not approving MR/PRs or doing the merging).

But still, this is very creative and a nice application of LLMs that isn't strictly barf.

MattyRad

9 か月前議論

From a quick Google query, it says that ~%90 of Americans have health insurance (which seems higher to me than I'd expected). I'd be very interested in knowing the number of uninsured, negligent/nefarious, and exorbitant invoices that are issued as a percentage of all invoices, for the purpose of determining the scale of criminality with respect to your description.

MattyRad

9 か月前議論

I glanced at Ubuntu Touch, but its device compatibility looked severely lacking (https://devices.ubuntu-touch.io/).... I have old Pixel phones I could potentially try it out on, but the last Pixel phone that is officially supported is the 3a. So that is a bummer.

投稿

Employers, please use postmarked letters for job applications (2025)

Ask HN: Have you naturally stopped using AI?

コメント