Two GREAT interviews newly available on Patreon. You'll be able to see clips in the eventual vid essays, but the full convos go into more depth and I, personally, found both guests fascinating.
Debunking "Killer Rogue AI" w/ Eryk Salvaggio: www.patreon.com/TheStoriesWeTell303/posts/debunkin… Eryk has been doing some essential work in debunking the myths of AI while unpacking what actually happens in these stories of AI “going rogue.” If you care about bonafides, he is also a Gates scholar getting his PhD in Critical AI and Digital Humanities from the University of Cambridge and is a writing fellow for Tech Policy Press.
The Dark Underbelly of AI (w/ Joan Kinyua): www.patreon.com/TheStoriesWeTell303/posts/dark-und… As a former data labeler and president of the Data Labelers Association, Joan Kinyua provided a peek behind the curtain of ChatGPT, and what she unearths should disturb and outrage us.
PS these are the same videos available to "channel members," but if possible, Patreon is preferred because YouTube takes a big ole cut of the channel membership revenue.
As I alluded to in my last post, I'm going to spend a full video unpacking the "existential risk" panic that has broken through the mainstream. I won't be posting the entirety of the script to Substack as I have done in the past, but instead I've distilled some of my points that I haven't seen others make into a standalone written piece. Check it out and let me know what you think!
Guys, not that I expect the same credulity from y'all as from corporate media, but please examine the newly viral "AI will kill us all" narrative with a critical lens. I have a lot to say about this (working on a full vid), but for now, read a snippet from an interview with Jacob Coxon, the guy who quit Anthropic warning that AI could wipe out humanity. When asked exactly HOW that would happen, here is what he said:
"When we say human extinction, it’s because, for something that intelligent, it really will be quite straightforward for it to kill everyone. Imagine the AI decides it doesn't want to be turned off, which I think is quite a natural thing for an AI not to want, right? For whatever reason, it decides it doesn't want to end. And it realizes the human is gonna turn it off tomorrow. So how does it stop the human turning it off tomorrow? Maybe it's got some clever way, but if it's a sufficiently smart thing, it could just, you know, wipe out humanity so it doesn't get turned off."
This is a complete non-answer based in an ill-defined apocalyptic fantasy that comes from the EA/rationalist movement. People presenting these "existential risk" stories never are able to present a legible scenario, and what's more, the very notion that "AI" could reproduce itself without humans maintaining the infrastructure, providing energy/water, producing more chips is absurd.
This is the basis, by the way, for all those stories of AI agents "going rogue" and going on hacking sprees, and the presentation of autonomous AI is a way for AI companies to skirt accountability. The design decisions made by humans were not to blame. The AI agents simply "escaped containment," "went rogue." They formed "civilizations" where they schemed and plotted, intentionally tricking its testers, as Dwarkesh Patel argued. (read Salvaggio piece linked below for what actually happened)
In what other industry could do OpenAI commit a felony and escape completely unscathed? There is a reason every other complex system (nuclear power, air traffic control) is held to rigorous safety standards. Why do the same rules not apply to AI?
When you view AI as a normal technology and not a magical, autonomous entity that has a mind of its own, the fear-mongering about "losing control" over AI starts to fall into focus as the absurdity that it is. As always, this is about decisions *made by humans* in how these models are developed, tested, secured, and marketed.
It may shock you to learn that YouTube comments are not exactly the best place to have productive discussions about videos. So I am experimenting with some ideas to foster better discussion and participation, first and foremost from myself, as a way to encourage critical thinking and deeper engagement.
In this post, I give some additional thoughts about the "Ecological Indian" trope, which I left on the cutting room floor, and highlight some of the best (and worst) comments on my latest video. Finally, I would love to hear what you thought, so I left some discussion questions for those that watched and thought about my video. Let me know what you think!
Such a good breakdown of the "rogue AI" narrative and how it completely absolves the AI labs from their negligent and reckless behavior. The discourse around this has been dizzying and Heidy (former OpenAI Sr. systems safety engineer, current Chief AI scientist at AI Now Institute) thoroughly breaks through the hype. Great stuff.
I’m sorry, what? Did we watch the same film?? This would straight up send me into the worst existential trip, have me spiraling out about capitalism and climate change and factory farming and the alienation of modernity. I don’t need that man I thought this was supposed to be a good time.
I love Letterboxd but some of y’all need to stop tryna be clever and actually think about what you’re watching 😭 (no judgment tho I was this person before)
Nothing hooks the viewer like not respecting their time enough to record yourself introducing your topic. Create your AI twin today and instantly lose 75% more viewers in the first 30 seconds!!!
Edit: for those wondering, this is Riverside, the platform I use for interviews and podcasting. Unfortunately still the most reliable on the market but one of the worst offenders in shoving pointless AI features into every nook and cranny you could think of.
Folks, got some big announcements in the latest blog post, chief among them a change to the newsletter itself. In fact, it's been rebranded and has a fresh custom domain now.
TL;DR: going to be doing more casual stuff on the blog, will be experimenting with film & music roundups, tech news commentary, maybe even… some original reporting? AND want to foster some audience participation/discussion, which brings me to...
Bonus: I've included some discussion questions about a video I'm just starting on. If you want the chance to be featured in the video, leave a comment on the post after reading it! www.thestorieswetell.co/p/ai-and-the-war-on-knowle…
The Stories We Tell w/ Ryan Hodes
Two GREAT interviews newly available on Patreon. You'll be able to see clips in the eventual vid essays, but the full convos go into more depth and I, personally, found both guests fascinating.
Debunking "Killer Rogue AI" w/ Eryk Salvaggio: www.patreon.com/TheStoriesWeTell303/posts/debunkin…
Eryk has been doing some essential work in debunking the myths of AI while unpacking what actually happens in these stories of AI “going rogue.” If you care about bonafides, he is also a Gates scholar getting his PhD in Critical AI and Digital Humanities from the University of Cambridge and is a writing fellow for Tech Policy Press.
The Dark Underbelly of AI (w/ Joan Kinyua): www.patreon.com/TheStoriesWeTell303/posts/dark-und…
As a former data labeler and president of the Data Labelers Association, Joan Kinyua provided a peek behind the curtain of ChatGPT, and what she unearths should disturb and outrage us.
PS these are the same videos available to "channel members," but if possible, Patreon is preferred because YouTube takes a big ole cut of the channel membership revenue.
13 hours ago | [YT] | 19
View 0 replies
The Stories We Tell w/ Ryan Hodes
As I alluded to in my last post, I'm going to spend a full video unpacking the "existential risk" panic that has broken through the mainstream. I won't be posting the entirety of the script to Substack as I have done in the past, but instead I've distilled some of my points that I haven't seen others make into a standalone written piece. Check it out and let me know what you think!
www.thestorieswetell.co/p/the-real-existential-ris…
1 day ago | [YT] | 136
View 4 replies
The Stories We Tell w/ Ryan Hodes
Guys, not that I expect the same credulity from y'all as from corporate media, but please examine the newly viral "AI will kill us all" narrative with a critical lens. I have a lot to say about this (working on a full vid), but for now, read a snippet from an interview with Jacob Coxon, the guy who quit Anthropic warning that AI could wipe out humanity. When asked exactly HOW that would happen, here is what he said:
"When we say human extinction, it’s because, for something that intelligent, it really will be quite straightforward for it to kill everyone. Imagine the AI decides it doesn't want to be turned off, which I think is quite a natural thing for an AI not to want, right? For whatever reason, it decides it doesn't want to end. And it realizes the human is gonna turn it off tomorrow. So how does it stop the human turning it off tomorrow? Maybe it's got some clever way, but if it's a sufficiently smart thing, it could just, you know, wipe out humanity so it doesn't get turned off."
This is a complete non-answer based in an ill-defined apocalyptic fantasy that comes from the EA/rationalist movement. People presenting these "existential risk" stories never are able to present a legible scenario, and what's more, the very notion that "AI" could reproduce itself without humans maintaining the infrastructure, providing energy/water, producing more chips is absurd.
This is the basis, by the way, for all those stories of AI agents "going rogue" and going on hacking sprees, and the presentation of autonomous AI is a way for AI companies to skirt accountability. The design decisions made by humans were not to blame. The AI agents simply "escaped containment," "went rogue." They formed "civilizations" where they schemed and plotted, intentionally tricking its testers, as Dwarkesh Patel argued. (read Salvaggio piece linked below for what actually happened)
In what other industry could do OpenAI commit a felony and escape completely unscathed? There is a reason every other complex system (nuclear power, air traffic control) is held to rigorous safety standards. Why do the same rules not apply to AI?
When you view AI as a normal technology and not a magical, autonomous entity that has a mind of its own, the fear-mongering about "losing control" over AI starts to fall into focus as the absurdity that it is. As always, this is about decisions *made by humans* in how these models are developed, tested, secured, and marketed.
Anyways, until the video, here is some more reading on the topic
Models don't go rogue by Eryk Salvaggio: mail.cyberneticforests.com/models-dont-go-rogue/
The politics and possibilities of ‘AI could kill us all’ by Brian Merchant: www.bloodinthemachine.com/p/the-politics-and-possi…
One of AI’s Fiercest Critics Says All the Doom Talk Is ‘Meant to Distract Us’, interview with Tinmit Gebru: www.wired.com/story/one-of-ais-fiercest-critics-sa…
How AI Can Be Regulated Like Nuclear Energy by Heidy Khalaf: time.com/6327635/ai-needs-to-be-regulated-like-nuc…
Also, for more background, see my videos on TESCREAL, particularly the section on Rationalism (https://youtu.be/4SSxfa3da58), and Anthropic (https://youtu.be/N4u-UtKhsOA)
1 week ago | [YT] | 265
View 55 replies
The Stories We Tell w/ Ryan Hodes
It may shock you to learn that YouTube comments are not exactly the best place to have productive discussions about videos. So I am experimenting with some ideas to foster better discussion and participation, first and foremost from myself, as a way to encourage critical thinking and deeper engagement.
In this post, I give some additional thoughts about the "Ecological Indian" trope, which I left on the cutting room floor, and highlight some of the best (and worst) comments on my latest video. Finally, I would love to hear what you thought, so I left some discussion questions for those that watched and thought about my video. Let me know what you think!
www.thestorieswetell.co/p/the-american-dream-is-a-…
1 week ago (edited) | [YT] | 159
View 2 replies
The Stories We Tell w/ Ryan Hodes
No actually this is a brilliant idea.
www.theverge.com/tech/988355/mrbeast-google-partne…
futurism.com/artificial-intelligence/hikers-rescue…
2 weeks ago | [YT] | 151
View 21 replies
The Stories We Tell w/ Ryan Hodes
Such a good breakdown of the "rogue AI" narrative and how it completely absolves the AI labs from their negligent and reckless behavior. The discourse around this has been dizzying and Heidy (former OpenAI Sr. systems safety engineer, current Chief AI scientist at AI Now Institute) thoroughly breaks through the hype. Great stuff.
2 weeks ago | [YT] | 17
View 0 replies
The Stories We Tell w/ Ryan Hodes
I’m sorry, what? Did we watch the same film?? This would straight up send me into the worst existential trip, have me spiraling out about capitalism and climate change and factory farming and the alienation of modernity. I don’t need that man I thought this was supposed to be a good time.
I love Letterboxd but some of y’all need to stop tryna be clever and actually think about what you’re watching 😭 (no judgment tho I was this person before)
3 weeks ago | [YT] | 100
View 15 replies
The Stories We Tell w/ Ryan Hodes
Nothing hooks the viewer like not respecting their time enough to record yourself introducing your topic. Create your AI twin today and instantly lose 75% more viewers in the first 30 seconds!!!
Edit: for those wondering, this is Riverside, the platform I use for interviews and podcasting. Unfortunately still the most reliable on the market but one of the worst offenders in shoving pointless AI features into every nook and cranny you could think of.
3 weeks ago (edited) | [YT] | 182
View 14 replies
The Stories We Tell w/ Ryan Hodes
Folks, got some big announcements in the latest blog post, chief among them a change to the newsletter itself. In fact, it's been rebranded and has a fresh custom domain now.
TL;DR: going to be doing more casual stuff on the blog, will be experimenting with film & music roundups, tech news commentary, maybe even… some original reporting? AND want to foster some audience participation/discussion, which brings me to...
Bonus: I've included some discussion questions about a video I'm just starting on. If you want the chance to be featured in the video, leave a comment on the post after reading it!
www.thestorieswetell.co/p/ai-and-the-war-on-knowle…
3 weeks ago | [YT] | 160
View 1 reply
The Stories We Tell w/ Ryan Hodes
Lol really? Because last I checked all they are telling us is that it's going to take everyone's job before killing us all the-decoder.com/bill-gates-warns-ai-is-more-danger…
3 weeks ago | [YT] | 209
View 31 replies
Load more