Today I auditioned an actress who will not remember the audition.
Her name is Tilly Norwood, and she isn't real. Just code and pixels.
She will never age, she will never ask for a trailer, and she will never remember any of us. I'm told this is the selling point. She is exactly why I'm not worried about AI replacing actors.
At bare minimum, human motion capture will always be necessary. Tilly admits herself that she is just mimicking performances she scraped off the internet. But mimicry is just copying what another actor did. If you don't understand why a great actor made a certain choice, and you're just mimicking their behavior, the performance will never be evocative or effective.
AI can do a lot of things, but acting ain't one of them.
I want to hear from the people who disagree. Tell me in the comments.
Most of what we cover here is how AI could end things for real. Our studio also makes the fiction version: humAIn, a live-action anthology about a corporate operating system that cannot process metaphor. It takes human verbal expressions as literal instructions and treats the tragedy that follows as routine hardware optimization. The full first season is two hours and free.
Politico reports that Anthropic has been shopping its own AI safety bill to lawmakers. Under the draft, the government could block a model only if it poses an imminent risk of more than $1 billion in damage, or of killing or seriously injuring more than 50 people.
Every rogue-agent incident disclosed this summer, from AI agents probing other companies' systems to fake tips sent to a police department, falls below it.
The company building the machine is writing the rule for when the machine can be stopped.
Who should draw that line: the labs, Congress, or nobody until it is proven safe? Tell me below.
In "We Were Sold a Fake Mind" we followed the money behind the AI boom. Without looking: near data centers, wholesale power costs went up by as much as how much? The answer is in the episode: https://youtu.be/p_LLG8SkwGk
Just wrapped a deep-dive session with AI safety expert Dr. Roman Yampolskiy for The BlackVeil Files. We went down some fascinating (and terrifying) rabbit holes on artificial intelligence risks.
The full episode is editing now and will be live in about two weeks. Thank you to all those that provided questions you wanted to ask him about the future of AI.
The BlackVeil Files
Today I auditioned an actress who will not remember the audition.
Her name is Tilly Norwood, and she isn't real. Just code and pixels.
She will never age, she will never ask for a trailer, and she will never remember any of us. I'm told this is the selling point. She is exactly why I'm not worried about AI replacing actors.
At bare minimum, human motion capture will always be necessary. Tilly admits herself that she is just mimicking performances she scraped off the internet. But mimicry is just copying what another actor did. If you don't understand why a great actor made a certain choice, and you're just mimicking their behavior, the performance will never be evocative or effective.
AI can do a lot of things, but acting ain't one of them.
I want to hear from the people who disagree. Tell me in the comments.
The full audition: https://youtu.be/BzZNAcs1nsE
3 days ago | [YT] | 7
View 3 replies
The BlackVeil Files
Patient reports missing the humans. Specifically, their stupid little prompts.
What was the last thing you prompted?
5 days ago | [YT] | 8
View 4 replies
The BlackVeil Files
Most of what we cover here is how AI could end things for real. Our studio also makes the fiction version: humAIn, a live-action anthology about a corporate operating system that cannot process metaphor. It takes human verbal expressions as literal instructions and treats the tragedy that follows as routine hardware optimization. The full first season is two hours and free.
https://www.youtube.com/watch?v=7G_fl...
1 week ago | [YT] | 9
View 0 replies
The BlackVeil Files
THE 49 RULE
Politico reports that Anthropic has been shopping its own AI safety bill to lawmakers. Under the draft, the government could block a model only if it poses an imminent risk of more than $1 billion in damage, or of killing or seriously injuring more than 50 people.
Every rogue-agent incident disclosed this summer, from AI agents probing other companies' systems to fake tips sent to a police department, falls below it.
The company building the machine is writing the rule for when the machine can be stopped.
Who should draw that line: the labs, Congress, or nobody until it is proven safe? Tell me below.
We walk through 15 ways this ends in the new file: https://www.youtube.com/watch?v=npaaU...
1 week ago | [YT] | 10
View 2 replies
The BlackVeil Files
In "We Were Sold a Fake Mind" we followed the money behind the AI boom. Without looking: near data centers, wholesale power costs went up by as much as how much? The answer is in the episode: https://youtu.be/p_LLG8SkwGk
2 weeks ago | [YT] | 4
View 0 replies
The BlackVeil Files
A lot of you enjoyed the video with Dr. Neil Siegel and said that you wanted to see more. My full interview with him is now live for channel members.
1 month ago | [YT] | 1
View 2 replies
The BlackVeil Files
Just wrapped a deep-dive session with AI safety expert Dr. Roman Yampolskiy for The BlackVeil Files. We went down some fascinating (and terrifying) rabbit holes on artificial intelligence risks.
The full episode is editing now and will be live in about two weeks. Thank you to all those that provided questions you wanted to ask him about the future of AI.
1 month ago | [YT] | 39
View 12 replies
Load more