Tar_Alcaran

Boys…

Nowadays, almost every hair iron has attachments for both straightening (with it a big clamp with two flat plates) and curling (various diameters of metal rod).

If you use a curling iron backwards, you burn your hands

Also pictured here: Anthropic stating out loud their models will just give out all the “secret” and “secured” internal data to anyone who asks.

Of course, that’s by design. LLMs can’t have any barrier between data and instructions, so they can never be secure.

Well, then for education, here’s what happens to any place without censorship.

Normal people are interacting happily. Some assholes come by and start being assholes without anyone stopping them. More assholes join because this is a great place for assholes. Normal people start getting annoyed with the great number of assholes and leave. It’s now a place for assholes.

For “asshole” fill in nazi, pedophile, manospherian, etc.

I distinctly remember the day I lost faith in educated people.

I do Workplace Safety and (waste) compliance. That day, I was at a rent-a-lab, telling people about how stuff works regarding their chemicals. Now, I have a doctorate in chemistry, but that room had more degrees than people by a pretty good margin.

I had to go over basic concepts like “the hand washing sink is not for rinsing chemicals. No, not even if you run the tap a lot” and “yes, you can run a waterhose down the special disposal sink, but it costs your company thousands of euros” several times.

Or “yes, if you scoop a little bit of [hazardous waste] into each bag of used napkins, you are technically below the level of it being a bag full of hazardous waste, but you’re also willingly and intentionally commiting ecological crimes and slowly murdering garbage workers. Yes, if I signed off on it, nobody would know, and that would make me an accomplice AND make this a conspiracy, so I suggest we don’t do things that will get me fined roughly all worldly possessions”.

Most damning was “no, these barrels aren’t magic. Anything you put in a barrel will stay in the barrel. Yes it says “disposal”, but they’re loose barrels. They don’t empty till someone empties them. Like a bucket”.

Actually, MOST damning was “no you can’t keep food in the lab. Not even in the special food-only fume hood”. But that was at a different place.

Everyone is on a list, the only thing that changes is the order of that list.

I’m a fucking ignorant moron on the vast majority of topics.

A dumb person will happily talk about everything as if they understand it. A smart person will shut about the things they don’t understand and be open to being educated.

Oh no, it’s actually much worse.

These LLM datacenters are basically useless for anything but running LLMs. You can’t use them for any other cloud software, because all the electronics were specifically made for running LLMs. Really, the only thing that will still be useful is the concrete building and the power and data lines coming in.

No, the actual reason why hundreds of billions of dollars are being spent is because of three things.

• AI boosters (Sam Altman, Dario Amodei, etc) are excellent liars and great storytellers, and rich people eat that shit up.

• rich people invest money by the bucket load, and everyone else notices. Stocks go up and everyone notices that LLM stocks go up, making people buy them making stocks go up because surely, something popular must also be good and useful.

• The people investing money into these lies want to see it happen, so they tell their pet CEOs to implement LLMs and reward them for implementing LLMs. Not for getting a use out of it, but for applying it.

So we get a bunch of people benefitting in the form of massive bonusses for implementing LLMs and pushing the use of LLMs, despite it actually not doing anything useful and being actively detrimental in most cases. All to make the stock line go up, unlocking more bonusses for the ultra rich.

But the LLM itself is pointless. It doesn’t even matter if it does anything. It’s just a vehicle to inflate fake numbers.

They kind of do. Not through some extra sense, but because they have insanely good hearing, and very low visual acuity combined with really good motion detection.

They have a completely different perception of the world, basically everything is a blur, but moving objects stand out. That’s why cats can stare at an unmoving laser dot: they’re not unmoving if you’re a cat.

I laughed at this so hard

Removed by mod

No, my house is definitely outdoors.

Yes, and that makes the plants in a greenhouse indoors. Or is an “indoor greenhouse” just a greenhouse placed inside larger house?

Isn’t a greenhouse by definition indoors?

Good. This brightens my day

Exactly. It’s like they don’t even realize who the protagonist is here.

Terms like “sense” and “tell” are a bit misleading. It’s very much a chemical/mechanical interaction that’s automatic. Rather like soap bubbles “sensing” when they’ve reached the surface of the water.

Plants contain a protein called phototropin, which is activated by light. When it’s activated, it changes the shape and alignment of the “skeleton” of the cell, making it more cube-shaped as opposed to long and skinny.

That means the light side of a plant gets shorter, while the dark side remains long. The dark side also grows slightly faster, on a count of having more cells there (you can fit more skinny cells side-by-side than wide cells), and so the plant angles and grows toward the light.

And yes. The colour matters. Phototropin reacts best to blue light, and leaves absorb mostly red and blue light (which is why they’re green). It basically ignores the green light filtered through leaves.

But that only kills those larva. Not the ones in all the ponds nearby.

I actually reached out to Cowan to asks a few questions. He was pumped that we were citing his work and using it in reclamation planning as landscape enchantments.

I’m in a completely different field, but there’s nothing more awesome than seeing your work get used in real life situations that actually match up with your goals.

And people showing a genuine interest is a close second.

It’s not that AI can replace them. It’s that LLMs can offer a hugely inferior version of what they do, and for a lot of people, that’s apparently fine.

Quite a few fields that previously had only the options of “good, quality work” and “not having it”, now have LLMs to offer the option of “substandard slop”.

And people like copywriters, translators, illustrators and other fields are rapidly finding out there wasn’t actually a huge demand for their good quality work, there was in fact a huge demand for “not nothing”, and that demand is being filled by the mediocre slop produced by LLMs now.