PauseAI•14mo ago

Emailing an influencer Penguinz0

Charlie White, otherwise known as MoistCritikal, is a bit of a goofy channel, but he's recently been expressing massive distaste for AI and its effect on jobs, which is great because he has 16.2 million subscribers and averages a million views per video. So I thought I'd send him an email asking him to raise awareness for pause AI. I was hoping people could check the email, and see if they could think of any way to improve it. FYI: the email is very similar to the one I sent my local politician, but the language is softer because (as a relatively famous individual) he's used to dealing with crazies.

then-purpleOP•11/19/24, 1:27 PM

I am writing to you today as a volunteer on behalf of Pause AI, an organization devoted to slowing down AI development to help mitigate risks such as economic inequality brought on by mass automation, AI generated misinformation, and AI enabled terrorist attacks. We were hoping that, considering your recent videos on AI, you'd like to assist us raising awareness about these issues, to educate the general public and encourage politicians to regulate AI. I understand you're a very busy person with many people asking for your attention, but a few moments of your time to read what I've written would be massively appreciated.
AI has many positive uses whether it's self-driving cars increasing safety on the road, advances in medicine, or battling the climate crisis; however, there are massive risks to the public that need to be mitigated in our view. Some of these issues may initially sound like science fiction, but they are taken very seriously from top academics, including Stephen hawking before his death.
A statement that was published by the centre of AI safety last year, signed by multiple Nobel prize winners, professors of computer science at Harvard, Oxford and Cambridge, the top three most cited AI researchers ever and hundreds more scientists of similar calibre, declared: "“Mitigating the risk of extinction from AI should be a global priority, alongside other societal scale threats such as pandemics and nuclear war.”
The science behind the problem takes three paragraphs to explain, I've attached sources at the bottom of the email.

then-purpleOP•11/19/24, 1:28 PM

AI is becoming incredibly intelligent, today they outperform 89% of programmers at coding (1), can design working rocket engines (2), and outperform PhD students in science exams (1). An AI called MegaSyn has invented 40,000 new possible chemical weapons (3) in six hours, some of which were similar to VX, the most potent nerve agent ever developed.
AI has an unfortunate habit of doing things we did not program them to do. E.g: we wanted to teach an AI to walk in a video game (4), so we gave it a pair of digital legs and tasked it to get from point A to point B. The scientists were surprised when it simply flopped on its side and started gliding from point A to B, it shouldn’t have been possible but the AI had exploited bugs the programmers weren’t aware of to achieve its goal in a way the scientists didn’t intend. It’s a trivial example but has loads of real-world implications favouring men over women in hiring processes (5), encouraging users to harm themselves (6), or social media algorithms causing political polarisation (7) and from this evidence we can conclude that AI, for the foreseeable future, will not be aligned with human values. For more examples click here.

then-purpleOP•11/19/24, 1:28 PM

Future AI may cause catastrophe, due to something called instrumental convergence, a law of nature which basically states “To achieve any goal you must not die”. You cannot make coffee if you’re dead, you cannot do maths if you’re dead, you cannot do backflips if you’re dead. For every goal the AI could conceivably have, it is obvious any sufficiently intelligent machine will take steps to ensure its own survival. From this we can predict that AI will exhibit certain behaviours, it will make copies of itself for redundancy, make sure it has access to a constant supply of electricity, and almost certainly prioritise its survival over human beings. We put kill switches in its computer chips, from this it will very likely surmise we’re a threat to its continued existence and attempt to get rid of us (8).

then-purpleOP•11/19/24, 1:29 PM

In short: super intelligent AI has the motive and means to cause massive loss of life.
Of course, there is intense debate about this and there are experts of high calibre like Yann Le Cun, Fei-Fei Li and Andrew Ng, who think the threat of human extinction is overblown. If you'd like to familiarise yourself with the arguments for and against, I recommend watching the Munk debate here. Overall, according to this survey about half of AI researchers believe there is a 10% chance or greater that AI causes massive loss of life.
As you can see we're very concerned about these issues, I'd be grateful and honoured if you could take a little of, I am sure, your very limited time, to discuss these matters and any possible solutions.
Sincerely Joseph Baker.

skinny-azure•11/19/24, 3:34 PM

I like it, but i'd need the links to check if they point to the correct thing. Otherwise no feedback, i also responded to the politician email and have nothing new to add

Sskinny-azure I like it, but i'd need the links to check if they point to the correct thing. O...

then-purpleOP•11/19/24, 3:57 PM

oh yeah, links are the same as last time, so links are good

sacred-rose•11/20/24, 5:47 PM

I like it! you should check out#Influencer outreach (e-mails).

dead-brown•11/21/24, 3:56 PM

Did you get a reply? It seems long to me

Ddead-brown Did you get a reply? It seems long to me

then-purpleOP•11/21/24, 5:14 PM

Ah nah, do you think I should send a follow up?

dead-brown•11/21/24, 5:20 PM

Yeah. Maybe more aggressive?

awake-maroon•11/27/24, 6:31 PM

Huge!

awake-maroon•11/27/24, 6:31 PM

Just write a few paragraphs next time, otherwise it's unlikely they'll read it.

awake-maroon•11/27/24, 6:32 PM

Almoast nobody wants to read a wall of text, and this person is busy, so it's important to keep it short with some space between the lines (some breathing room)

constitutional-coral•11/28/24, 10:41 AM

I watch a lot of his content, ping me know if you need help.

sad-indigo•11/28/24, 11:14 AM

I want to help

Cconstitutional-coral I watch a lot of his content, ping me know if you need help.

then-purpleOP•11/28/24, 12:02 PM

Help as in also sending emails or editing emails?

then-purpleOP•11/28/24, 12:03 PM

Also here’s the follow up email I wrote: I hope this message finds you well. I wanted to follow up on my previous email regarding thePause AI initiative. I understand that you're likely juggling many important projects and responsibilities, but I believe the issues we've discussed are critical and warrant your attention.
As I mentioned before, the rapid advancement of AI presents both incredible opportunities and significant risks. The potential for AI to outperform humans in various tasks is undeniable, but so are the risks associated with misalignment, unintended consequences, and the potential for catastrophic outcomes. The concerns aren't just theoretical—they're backed by some of the brightest minds in the field.
Given your influence and the reach of your content, your voice could play a pivotal role in raising awareness about these issues. Mark Ruffalo and Joseph Gordon-Levitt have already highlighted the importance of AI safety, and your support could further amplify this critical message. Additionally, J.J. Abrams, Mark Hamill, Pedro Pascal, and 125 other celebrities signed a letter to the Governor of California, urging for basic AI safeguards. Their collective action underscores the importance of influential figures speaking out on this subject. I truly believe that together, we can make a meaningful impact in ensuring that AI development is guided responsibly.
I would be honored to discuss this further with you and explore how we might collaborate to promote awareness and encourage regulation. Your insights and perspective would be invaluable.
Looking forward to hearing from you.

constitutional-coral•11/28/24, 12:04 PM

editing I guess. I don't really know how to reach out to him other than email, unless we find a way to contact his moderators or comment live twitch streams where he is very attentive.

Cconstitutional-coral editing I guess. I don't really know how to reach out to him other than email, u...

then-purpleOP•11/28/24, 12:07 PM

Yeah I was thinking about his streams as well, but I’ve so far missed them. Maybe we should keep an eye out, if I spot one I’ll message you about it, and vice versa, then we can try and get the message across? I’ll try and write down what we could send him

then-purpleOP•11/28/24, 12:07 PM

Or is that a bad idea?

constitutional-coral•11/28/24, 12:08 PM

I dont join his streams much because of time difference but it could work. you have to buy subscription so that is an issue.

constitutional-coral•11/28/24, 12:09 PM

He streams like everyday so its not an issue.

constitutional-coral•11/28/24, 12:10 PM

One thing about him though, if he does get interested he could possibly donate a significant amount. He is a decent person and does giving away money streams, plus a big name in YT community.

constitutional-coral•11/28/24, 12:19 PM

https://www.twitch.tv/moistcr1tikal/schedule here is his streaming schedule I don't see anything, in next few days which is strange,

Twitch

moistcr1tikal Schedule - Twitch

Check out moistcr1tikal stream schedule, and set reminders so you don’t miss out!

Tthen-purple Also here’s the follow up email I wrote: I hope this message finds you well. I w...

awake-maroon•11/28/24, 1:41 PM

Looks decent!

Potential improvements: stronger hook, have the hook earlier, what does this person gain, shorter, more breathing room (can't be a wall of text)

then-purpleOP•11/28/24, 1:52 PM

Yeah, that’s something I’m kind of struggling with. I think my worst fear is the email just comes off crazy without elaboration. How would you open it?

constitutional-coral•11/28/24, 3:00 PM

The second para isn't needed, he knows it's not great for us. you could likely move the signing of the letter to be the hook.

constitutional-coral•11/28/24, 3:01 PM

Also, wouldn't it be better to be reaching out on behalf of PauseAI and specifically mentioning the protests and actions to reiterate what he can expect from the organisation.

Tthen-purple Also here’s the follow up email I wrote: I hope this message finds you well. I w...

sad-indigo•11/30/24, 12:22 PM

Cool

Emailing an influencer Penguinz0

Similar Threads

Similar Threads

Similar Threads