NASA doesn't have enough radioactive material for its current needs, RTG is used only for missions far from Sun (and Earth).

I think it counts as necromancy. After all it brings frozen hamsters back to life.

Amazon's Mechanical Turk exists since 2005, so we are 20 years in the future

MechanicalTurk is for desk jobs and for tasks that originate as ideas in a human mind -- even if they get routed via an API.

Here we are talking about AI agents coming up with a set of tasks as part of their thinking/reasoning step ..and when some of those tasks are real world physical tasks, assign them to a willing human being.

Those tasks wont necessarily be desk jobs or knowledge work.

It could be say -- go chop a tree, or go wave a protest banner, or go flip the open/close sign on my shopfront, or go and preach crustafarianism.

Mechanical Turk was for humans to rent a human, which is not a new idea

mTurk has an API (and I guess it had it since the beginning). It is, of course, very AWS-que, but LLMs should be able to use it just fine.

∗ ∗ ∗

> which is not a new idea

I don’t think “[x] but for agents” counts as a new idea for every [x]. I’d say it’s just one new idea, at most.

I mean, the entire name of Mechanical Turk plays on "packaging up humans as technology", given the original Mechanical Turk was a "machine" where the human inside did the work.

It is almost like finding 20 year old bugs on Mozilla tracker. That said GitHub doesn't have the excuse of mostly relying on volunteer work.

Also I don't find GitLab that much better. I remember the feature request for "Give option to disable automatic adding of 'Closes ISSUE' to merge requests" closed with "Why would you need an option for that, everyone either loves it or likes manually removing it every time.

I would love AI to take over monitoring. "Alert me when logs or metrics look weird". SIEM vendors often have their special sauce ML, so a bit more open and generic tool would be nice. Manually setting alerting thresholds takes just too much effort, navigating narrow path between missing things and being flooded by messages.

I still think you're going to be in manual threshold tuning for quite a while. The cost of feeding a continuous log to an LLM would be insane. Even if you batched until you filled a context window.

Sending screenshots of charts and dashboards is also effective, and often context-window-friendlier

> So why are all the official sources still so paranoid about Vitamin D

It is fat soluble vitamin, together with A, E and K. That in itself makes in more risky in terms of overdose. I didn't hear of any cases outside kids eating jars of vitamin gummies but it does happen.

No, it reads like Linkedin post. That said, do we now have to check if the text we wrote doesn't look like something AI generated?

You're absolutely right.

Um, I got called on HN three times now accused of being AI for writing comments by hand.

I got so annoyed at the second time that I even created a post about it. I guess I just get really annoyed when someone accuses me who writes things by hands as AI slop because it makes me feel like at this point, why not just write it with AI but I guess I just love to type.

I have unironically suggested in one of my HN comments that I should start making the grammatical mistakes I used to make when I had just started using HN like , this mistake that you see here. But I remember people actually flipping out in comments on this grammatical mistake so much that It got fixed.

I am this close to intentionally writing sloppy to prove my comments aren't AI slop but at the same time, I don't want to do this because I really don't want to change how I write just because of something what other people say imo.

Don't kid you'reself, people LOVE grammatical and spelling errors. It's low entry, and by far the easiest way to get someone to interact with what you have written.

AI deprives them of this.

Why even read something with no mistakes? Just scan on to the next comment, you might get a juicy "your/you're" to point out if you don't waste time reading.

That's EXACTLY what an AI member of this community would say!

I know you're secretly a bot, because you used punctuation. Only AI uses punctuation!

/s

xD

that /s is carrying the whole message haha

but yeah I guess, sometimes I wonder if suppose a bot was accused of being AI, I mean if trained with right prompt and everything, it can also learn to flip out and we would be able to genuinely not trust things.

I guess it can be wild stuff but currently I just really flip out while literally just being below the swear level to maintain decency (also personally I don't like to swear ig) to then find that okay I am a human after all.

But I guess I am gonna start pasting this youtube video when somebody accuses me of being AI

I am only human after all: https://www.youtube.com/watch?v=L3wKzyIN1yk

It would be super funny and better than flipping out haha xD

"Got no way of prove it so maybe I am lying but I am only human after all, don't put the blame on me, Don't put the blame one me" with some :fire: emoji or something or not lmaoo. It would be dope, I am now waiting (anticipating out of fun) for the next time when I comment something written by me (literally human lmaoo) and someone calls me AI.

The song is a banger too btw so definitely worth a listen as well haha

It really is nothing new. People quickly close windows with errors, they go out of they way to avoid reading actual message.

Not even close. The 2008 financial crisis is better comparison. And even then I think most negative effect will come from investors pulling money away from everything non-AI, than OpenAI/Anthropic/Oracle crashing and burning.

I think the only reason left, that we don't have to use ML to detect ads in YouTube streams, are the legal requirements for visible separation of content and ads. I doubt LLM ads will get more integrated than current Google search results screen. Maybe services that don't have enough "surface" for ads (including all APIs) will move to subscription only model.