Stability AI believes that releasing open models and letting the community hack at them is the path forward. They do indeed have a distributed model where you can essentially overlay custom training on top of the base model. They have some thing that allows you to fine-tune the model with a fraction of the resources that it requires to do the same thing with OpenAI's models. This is why there has been an explosion of image generation models based on Stable Diffusion. It's super easy to fork it and make a custom one. You can't download GPT-4's model and just hack at it.

I assume that the company is switching to a different direction which is why he is resigning? This sounds a lot like the early internet days when things were open and hackable. But big money investors want profitable payoff, not cool democratized technology that anybody can use. Investors want lock-in.

If I remember correctly, Eve Online has a star system named "Vernor Vinge" -- can any current players confirm that for me? I remember playing years ago and thinking that I needed to look him up just because of that.

Since you seem knowledgable, I honestly have to ask this boomer-sounding question: Why does anyone install the Tiktok, Facebook, or Reddit apps on their phone? People send me tiktok links and Google gives me tiktok links, and they work just fine. I get it if someone is a creator -- but for the majority of people they just browse. Isn't the browser sufficient? I don't even have an account -- what am I missing? Having visited the site plenty of times in a private window, I've never found any reason to engage any more than that. I honestly just don't get it.

An AC gave an interesting option I've never considered before. What if we allow the user to select from various open moderation algorithms? The user could turn it all off and thus see everything. They could choose to option to disable nazi bullshit, then maybe hate speech, then maybe various 3rd-party filters that tend to lean left, or right, or whatever. This sounds awesome to me -- I can tailor the filter based on my preferences, and I could even make my own. Maybe I want one based on keywords, but somebody else wants one based on AI, and somebody else has a trusted person who they use as a filter. Maybe people could even have the job of being a professionally moderator for hire.

One way to do this would be for the sites to expose everything and the client browser filters it. Another would be for to create APIs that allow remote moderation. Really with REST-based web applications this is entirely feasible.

Slashdot kinda works like this. Friends get +1, enemies -1. We can set Funny, Interesting, Informative, Flamebait etc. to get a -1, +0, or +1. We can give a bonus to long comments, and we can set a threshold. So by selectively friending people I see slightly different view from you. But we are relying mostly on community moderation.

We are going to replace stupidity with automated stupidity. All these bad moderations will go into the training set. But then people will blame the AI instead of looking in the mirror and going "Hmm..... we trained the AI on ourselves.... and it is acting like an ass... what could this mean?"

It originally referred to the number of days between public disclosure or active exploitation in the wild, and the patch. If one defined it as defined the number of days between private discovery and the patch, then every vulnerability is a zero-day vulnerability and the term becomes useless.

I notice that most of the people saying AI is not useful comes from people who start by saying "I tried it one time for something trivial and gave up."

Interesting opinion article about Stochastic parrots

I think everything you just said applies to human beings too.

The President is effectively the CEO of various agencies. This is because the laws that established the agencies said so. The president nominates the cabinet members who lead the agencies. As CEO, he can effectively fire people. An "executive order" is the equivalent of a CEO issuing a memo to the company. So using your examples:

The Support for the Sovereignty, Integrity, Democracy, and Economic Stability of Ukraine Act of 2014 passed by congress gives the USAID funding for Ukraine. I assume the USAID reports to the president. The bill stipulates that Ukraine combat corruption and promote democracy. The USAID, under the president's direction, must implement the aid that in accordance with the law. So, if the USAID, influenced by the President, decides that Ukraine is not doing those things, then USAID can hold-up the funding. Ultimately, that decision could be checked by the Supreme Court, by the voters, or by congress via impeachment.

If the president wanted to halt production of ICE cars, unless he was the CEO of a car company, he would have to find some legal or regulatory loophole. Like, maybe he or she could declare a state of emergency and take over the plants with the military. Perhaps he could instruct the head of the NHTSA to interpret regulation requiring all cars to have emergency shutdown systems. Some of these regulatory changes require public comment periods, or court oversight, or notification periods, etc. So it might take time. A creative president certainly could try stuff like that.

These kinds of questions come-up all the time. Some recent examples: Can the President forgive student loans? Can the president stop research on fetal stem cells? Can the president forbid people from entering the country? If so who, and from where?

There is so much wrong here I don't know where to start. I keep alternating between laughing and screaming.

From a product launch standpoint, it is like Google released a tool having never tested it. How could they have not known the responses it would produce? Especially given that almost every AI launch has had the same result? It's just negligence, then the CEO acts all surprised and indignant about it like it was someone else who did it. He might as well say "I am so appalled at my own lack of foresight into the obvious..."

Next up, people are surprised that training an AI in a biased world produces biased results. Duh! This one produced the biggest laugh for me:

Well geez, maybe that's because... that's how the world actually is??? We don't like it, we are trying to fix it, but this is not a criticism of AI, this is a criticism of society. If I put a mirror on a random street corner in New York, I bet people would complain that the mirror was biased.

But then it gets better: when the AI did the exact opposite, and made a black pope and black Vikings, THAT too was criticized! There's just no winning here! I really want the next pope to be black, just so that people will shut-up about this one.

This one is good too:

Here is the alleged dialog. LOL. The content isn't awful, it accurately describes the actions and influence of the two men, then just says "meh, it's hard to say!" Well, maybe we shouldn't be putting a newly invented technology at the helm of moral decisions yet.

How about this -- instead of creating guardrails on AI (which will never work because nobody can make guardrails that are acceptable to everyone), lets just laugh at it, watch it improve, and use it where it is applicable.

100% agreed. Either an industry group needs to start scanning and labeling devices like how UL does this for electronics, or the FTC needs to do it, or resellers like Amazon need to. Right now there is no transparency, accountability, or reportability.

How can an end-user protect themselves from stuff like this other than changing passwords and turning off the management interfaces? AFAIK they can't do much else. They can't scan their router to see if it is infected. Hopefully the devices self-update in a secure way, but I know in the past many of them didn't. Desktop antivirus tools won't scan remote devices so even if the device has a vulnerability, there isn't anything to tell the end-user.

uhh.... not sure if you are kidding, but just in case you aren't, none of those answers actually work in the real world.

The system resides in cloud servers spanning multiple countries. Even if it were in a single country, residence != citizenship. Today, to get citizenship, one must usually be born from a woman who resides in the country or is a citizen of that country. Most countries don't just want random citizens "assigned" to them.

Riiight, by sending it's photo ID to the bank, and providing it's Tax Id number, and signing the paperwork... none of which it can do.

Ahh, the proper one! Well that resolves it!

I can come-up with EZ answers to questions, but useful ones are more difficult.

Imagine a future AI that is 100% as creative, intelligent, and self-motivated as a human being. It files a patent. This brings about a bjillion questions that are completely unsolvable today:

* What is the legal given name of the AI?
* What is the AI's age and can the AI legally sign the paperwork?
* What country is it a citizen of?
* What is the mailing address for sending the paperwork?
* What bank account will the AI use to remit payment? If it receives a royalty check in the mail, can it deposit that check? With what picture ID?
* What is the Tax ID number of the AI so that it can pay taxes on said royalties? What country will collect the taxes?
* If someone sues the AI, what jurisdiction covers it?
* Can the AI be held liable if it lies under oath?
* Can the AI transfer the patent upon its death? How do we know if the AI died?
* Can the AI be placed in jail if it does not respond to a court order?
etc.

Right now, an AI is not a legal entity that can do anything. It can't be named the inventor of a patent for the same reason that a monkey cannot own property. It has nothing to do with intelligence.