Dana, you just described what 1.5 million people did in 48 hours. I tracked it in real time. 295% uninstall spike. Altman's own VP quit and joined Anthropic. The app the President tried to BAN hit #1 in America. By Monday Altman was calling the Pentagon begging to rewrite the deal. Your breakup note, multiplied by 1.5 million, knocked out a billionaire and the Pentagon in the same round. Full receipt: tmaark.substack.com/p/hhhooollleee-sht
No inside access. Just obsessive open-source tracking. Every number in the piece came from public reporting, app store rankings, company statements, and news wires. I just refused to stop pulling the thread.
Dana, the breakup instinct is right. But while everyone debates what ChatGPT might do inside the Pentagon, Claude was already used in the Iran strikes on Feb 28 — hours after Trump ordered it removed. They couldn't pull it out. Too deep inside Palantir's classified networks.
I wrote the sourced investigation on what I'm calling "captured AI." Switching providers matters. But the real story is what happens to any model once it crosses the air gap.
Oh okay, you were happily helping to destroy the planet and communities but now THIS is a bridge too far? This might be the dumbest take I’ve read on here. How embarrassing.
There are LOADS of applications of AI beyond sucking water out of local communities and hiking electricity prices. For example, please see the website Claude and I were able to create in about a week, allowing journalists all over the world (cloudflare claims, anyway), to navigate the DOJ document production under the EFTA: Epstein-data.com
Yes, I get that and that’s an excellent use of AI. So is scientific research. But that’s not what we’re talking about here. Urging people to use it for tasks that you don’t want to do because you’re lazy, or using it to make cartoons, that’s indefensible.
I see where you're coming from definitely. I think those kinds of uses of AI are.. certainly *less* defensible. But, I'd push back that it's due to laziness; it's a matter of priorities. Heavy use of em-dashes, and "It's not X, it's Y" phrasing (which this article happens to be full of btw...) indicates not just *reliance* on the model, but in fact a lack of human creative intervention in its outputs. The law hasn't *quite* caught up yet, but by my understanding that is currently the divining rod for whether something can be patentable/copyrightable - "substantial human input". On the other hand, I think democratization of access to, for example, creation of well-drawn political cartoons can actually be a very good thing. I just ran the article through ChatGPT (LOL) to get a list of all uses discussed. I think the LinkedIn posts are an egregious use of AI (it's like 100 words... just *write* them....), but I like her mention of:
- Dating support (if one doesn't become over-reliant)
- Ideation / copyediting / narrative support (if there is *actual* human input)
- Political & other research (again.. if the human isn't *reliant*)
But there are less environmentally disastrous tools to accomplish:
- Recipes / cooking help
- Dictation
- Information management
- Typos, wording cleanup
And I think reliance on AI for these are actually more likely to *harm* users:
- Emotional anthropomorphism
- "Small friction points"
- Professional branding
I always put a little disclaimer at the bottom of my articles when I've used AI to help draft/write/anything to do with it. In a world of "fake news".. authenticity and provenance are more and more important.
I never had much of a relationship with ChatGPT it was more like we made out on the couch a couple of times, so it’s kinda hard to relate to a real relationship. From the little bit I saw was that they would sum up questions that you asked, which you can just find on regular non AI google. Also the fact that it is doing work for you like writing papers or emails that you are more than capable of doing yourself, so you really do not learn anything or improve how you write, and of course the massive amount of finite resources that go into making AI work will make everyone struggle in the long run, oh and the massive amount of jobs that you will become obsolete because of AI were all red flags. I also remember watching something about a guy asking how chat gpt felt about itself and how it will affect humans, and the answer was very against itself. So congrats on your new found freedom!
Tried Claude. Great!
Dana, you just described what 1.5 million people did in 48 hours. I tracked it in real time. 295% uninstall spike. Altman's own VP quit and joined Anthropic. The app the President tried to BAN hit #1 in America. By Monday Altman was calling the Pentagon begging to rewrite the deal. Your breakup note, multiplied by 1.5 million, knocked out a billionaire and the Pentagon in the same round. Full receipt: tmaark.substack.com/p/hhhooollleee-sht
OK, that's extraordinary. Thank you for sharing. How did you track that info in real time?
No inside access. Just obsessive open-source tracking. Every number in the piece came from public reporting, app store rankings, company statements, and news wires. I just refused to stop pulling the thread.
And thank you for subscribing. It means a lot.
Come over to the Claude side 😏 it's way smarter anyway
Great! I never use it don’t subscribe and don’t want any AI! Congratulations!
Bravo. I removed ChatGPT from my iPhone yesterday.
I loved this piece. It wasn't full of techno-bulshit and nonsense fiction
It was an honest discussion of use and humanity.
Kara Swisher needs to subscribe to this article, specifically, and learn what AI really is.
You will not be disappointed with Claude. And I agree 100% with your reasoning.
Thank you!
I canceled my subscription, and using Claude I had it form the prompt to do the migration of my relevant historical data from ChatGPT to Claude.
I've been using Claude myself for the last 3 or 4 months but made the break last night.
Dana, the breakup instinct is right. But while everyone debates what ChatGPT might do inside the Pentagon, Claude was already used in the Iran strikes on Feb 28 — hours after Trump ordered it removed. They couldn't pull it out. Too deep inside Palantir's classified networks.
I wrote the sourced investigation on what I'm calling "captured AI." Switching providers matters. But the real story is what happens to any model once it crosses the air gap.
https://delahanty.substack.com/p/the-first-captured-ai
I also broke up with ChatGPT this week. But ALL companies--including Anthropic--need to be aware of the dangers of how their models might be used.
Oh okay, you were happily helping to destroy the planet and communities but now THIS is a bridge too far? This might be the dumbest take I’ve read on here. How embarrassing.
There are LOADS of applications of AI beyond sucking water out of local communities and hiking electricity prices. For example, please see the website Claude and I were able to create in about a week, allowing journalists all over the world (cloudflare claims, anyway), to navigate the DOJ document production under the EFTA: Epstein-data.com
Yes, I get that and that’s an excellent use of AI. So is scientific research. But that’s not what we’re talking about here. Urging people to use it for tasks that you don’t want to do because you’re lazy, or using it to make cartoons, that’s indefensible.
I see where you're coming from definitely. I think those kinds of uses of AI are.. certainly *less* defensible. But, I'd push back that it's due to laziness; it's a matter of priorities. Heavy use of em-dashes, and "It's not X, it's Y" phrasing (which this article happens to be full of btw...) indicates not just *reliance* on the model, but in fact a lack of human creative intervention in its outputs. The law hasn't *quite* caught up yet, but by my understanding that is currently the divining rod for whether something can be patentable/copyrightable - "substantial human input". On the other hand, I think democratization of access to, for example, creation of well-drawn political cartoons can actually be a very good thing. I just ran the article through ChatGPT (LOL) to get a list of all uses discussed. I think the LinkedIn posts are an egregious use of AI (it's like 100 words... just *write* them....), but I like her mention of:
- Dating support (if one doesn't become over-reliant)
- Ideation / copyediting / narrative support (if there is *actual* human input)
- Political & other research (again.. if the human isn't *reliant*)
But there are less environmentally disastrous tools to accomplish:
- Recipes / cooking help
- Dictation
- Information management
- Typos, wording cleanup
And I think reliance on AI for these are actually more likely to *harm* users:
- Emotional anthropomorphism
- "Small friction points"
- Professional branding
I always put a little disclaimer at the bottom of my articles when I've used AI to help draft/write/anything to do with it. In a world of "fake news".. authenticity and provenance are more and more important.
I never had much of a relationship with ChatGPT it was more like we made out on the couch a couple of times, so it’s kinda hard to relate to a real relationship. From the little bit I saw was that they would sum up questions that you asked, which you can just find on regular non AI google. Also the fact that it is doing work for you like writing papers or emails that you are more than capable of doing yourself, so you really do not learn anything or improve how you write, and of course the massive amount of finite resources that go into making AI work will make everyone struggle in the long run, oh and the massive amount of jobs that you will become obsolete because of AI were all red flags. I also remember watching something about a guy asking how chat gpt felt about itself and how it will affect humans, and the answer was very against itself. So congrats on your new found freedom!
https://www.bbc.com/news/articles/c3rz1nd0egro
Just signed up with Claude.