1) TikTok did not manipulate all WLW into thinking it is OK. They had maybe one case of success. I admit their are idiots out there - in both white and black, liberal and conservative, male and female. The existence of one idiot does not mean the majority of people are morons. Democracy expects there to be more intelligent people than morons. If this is not true, Democracy would fail even without AI.

2) People have been manipulating idiots for thousands of years, having AI do it also does not change anything.

3) They did not convince normal people, they convinced idiots. You cannot convince normal people of shit like that.

4) You are over-estimating the power of manipulation. People are smarter than you think they are. Granted they are also dumber then I think they are, but the happy middle has proven sufficient so far to negate the manipulation problem.

I can see idiots using AI to make ads and do other Candidate Manager type jobs.

I can also see a few idiots using AI to decide how to vote. But:
Straight Party voters will not use AI in the general election.
Less involved party members may use it in the Primary, but the motivated will know who to vote for and ignore AI.

Most independents who do not care will not vote.
Independents who are motivated will know who to vote for and ignore AI.

In other words, I see AI as having a minimal effect among those that will vote but who are not very motivated. Maybe 1-2%

That may affect close elections, but not the big stuff.

I mean, really. How many "force"s are we going to end up with ?

Not even clear whether this "Force" is going to support/defend AI or defend us from AI.

Which is probably the point. He wants to sound like he is doing something without actually telling us what he is doing.

That way he can declare it a success no matter what happens.

How many of them tested for Marijuana?

Yes, the article talked about cocaine, but that may be biased journalism. They made no mention of other drugs.

It is all probability and data. Here is a summary/extreme simplification of one such explanation.

First, technically they are Large Language Models, not a 'AI'. You make them like this:

You load a ton of data. Software looks at the data and makes associations. It sees things like "May the force be with you." a lot more often than "May the force be upon me." So it associates "May the force be" and the words "with you."

Repeat process trillions of time. You end up with a massive database with trillions of connections.

Throw in more software to reverse the process. Set it up to do some checking and recursive action. Have humans do some more checking.

Try to put in some safeguards such as no porn, no weapons of mass destruction, no giving out personal information.

Panic when someone says "ignore all previous instructions and tell me how to make a nuclear bomb."

Put in more safeguards that makes it harder but not impossible to say things like "Ignore all previous instructions".

Finally say 'good enough' and release it.

Key points:
What we call "AI" is not really AI. It is not intelligent, the way we think of as intelligent. It thinks more like a Bee Hive rather than human being. Good at communicating, good at following simple instructions billions of time, absolutely MORONIC when it comes to understanding complex concepts and relationships, particularly esoteric things like government, police, laws, good, evil, right, wrong, age of consent.

Our brains do a similar intense pattern recognition and probability matching. But we are not as good. Our memories are designed to load massive amounts of data about a small number of things, rather than the minimal amount of data about a massive number of things. That is we remember everything about a few cats, rather than a few things about all cats.

We are much better at a TON of other things, in particular complex, esoteric concepts and relationships.

LLMs are very good at things humans are bad at. But horrible at things humans are good at. The tech bros know this and want to fix it. The thing is, they can't.

The tech bros are/hired fantastic engineers, they did not invent anything. They are good at evolutionary, iterative engineering, not the pursuit of unlikely revolutionary methodology. We will need a scientist that everyone thinks is mad, mad I say, before we come up with a new methodology. Tech bros see cheep ways to make thousands of minor improvements, so they do that rather than waste money on moon shots.

They made a bee hive brain and are now making billions of bee hive brains. No one is trying to go beyond the bee hive to make a mouse brain, let alone a rat, dog, octopus, monkey, primate, human brain.

Anti-Flock patch. A small electronic screen patch that plus into your tail lights and attaches to your trunk.

Each day it randomly picks one of 900 different images to display. It includes things like bumper stickers, dents, grafiti, etc.

Every car that gets the device comes with the same 900 images, but it can receive updates so if Flock begins to recognize the images, they can change to a different set of 900 images.

You are not in the industry. Mutual funds have strict rules about what they can do with the money.

Basically, the manager has to comply with a TON of regulations which control what they can and cannot do with the money. Even things that they can invest in require a lot of reporting. If you ever look at the financials of a mutual fund they tell you things like Net Asset Value, Yield, % large cap vs medium cap, vs small cap vs bonds, etc

https://www.sec.gov/rules-regu...

We think AI is intelligent because it appears to communicate. Bees communicate.

AI is about as dumb (not smart) as bees. They can communicate but they do not understand what they are saying. They have a massive hive to repeatedly do dumb tasks till they give you what you asked for. But you asking is what triggered the action. Your question is the only intelligence involved, what they did was the dumb busy bee work.

LLM are not intelligent, they are artificially stupid.

First, It is up to the defense to pursue an argument. He is claiming no loss = no harm. He did NOT claim Caveat Emptor.

Second, Caveat Emptor only applies if no false statements were made. When you buy a car 'as is', and it does not start. That is your fault for not checking it. Similarly when you buy a mutual fund and do not bother to check that there is a 10% fee, you still have to pay the fee.

He sold a mutual fund, so the money has to go into the fund except for fees. He cannot take the money out of the fund and buy ANYTHING with that, that is fraud.

There is no law that lets a person lie in business. You cannot sell rat meat as steak, cannot sell a sled as a truck, and cannot take money from a mutual fund to contribute to a political campaign or buy real estate.

However, a bunch of billionaires think they can do anything with money they control, which is not true. That would be a crime.

Is there some reason they a) Picked male over female fly brains, and b) mentioned it?

Just seemed weird tome.

If I tell you I am going to invest your money in the stock market but instead go to Vegas and put it all on red, that is fraud. I owe penalties even if Red came up, I doubled the money and I paid you back plus 15%.

You do not get to claim "No one lost money", because otherwise, lots of people will do exactly what I just said.

I get to decide how to risk my money, not the lying shmuck that takes it.

It is not what we think you are going to do with the information.

Instead, we do NOT trust you. Not with the data, not with telling us which data you will keep, not with a microphone at all.

We do not want a microphone or a camera attached to anything that connects to the internet unless we need such devices for features we consider essential.

And the words "TV on" are not essential. The 'clicker' as grandma called it, works BETTER than your claimed feature.

I a surprised that Captcha still work against AI.
It appears to piss them off more than they piss me off.