I feel as if some sort of government/AI company showdown was always going to happen because governments usually don’t like it when private entities and individuals control powerful, humanity altering technology. In a normal administration, this would
probably involve a long drawn out process of reigning in and regulating tech companies (or not). Instead, they went all in for a man they believed would de regulate their industry AND leave them alone to do as they pleased.
“Simply put: Claude can make mistakes, and when Claude is used in applications where a mistake has the potential to irrevocably change existing human lives, a human or company should be accountable for those mistakes.”…no truer words have been spoken than these. It irritates me to no end those who claim that the liability would be too large, and subject to regulations in support of this, it would stunt the development of these “life changing” technologies. To that I say, “good!”, perhaps it would force them to look deeper and work harder to make sure these AI technologies are well understood, predictable, and controllable. Instead we get their claims that it could doom society, but they’ll keep moving things right ahead, as though they have no choice but to proceed irrespective of the potential dangers that they themselves acknowledge.
As for this predictable outcome from Hegseth, that was telegraphed with Trump’s EO and his appointment of tech utopian David Sacks as Czar of tech pipe dreams.
Dear Dr Casey Mock, you begin by pitting the extremists back to back, the techno-optimists (accelerationists) and the Cassandres (doomers), to which I belong, along with my former PhD supervisor in Montréal. Admittedly, that's easy!
Then, in the rest of his post, you attack the Cassandres and proposes a more reasonable and balanced approach, arguing that AI is a technology like any other, useful but nothing more. You essentially criticize the Cassandres for highlighting the power of AI technology, which attracts power-hungry individuals and "dark triad" types (egocentric, manipulative, psychopathic) like the Secretary of War.
On this point, you're probably wrong. Trivializing AI is like comparing the invention of dynamite to that of the atomic bomb. There are orders of magnitude of difference and potential negative impacts. The same is true for those comparing the use of calculators and generative AI in education. Not to mention the effect of intellectual laziness on our brains.
AI has the potential to change the world, but it is up to us, Humans, to democratically decide which direction to take in the best interest of all, not just a handful of technobillionaires.
Even if the Cassandres "would try to outsmart" by trivializing AI, die-hard techno-optimists like Sam Altman, who promise nothing less than "eternal life" and an "infinite cornucopia" (sic, in a finite world), will have convinced the power-hungry ones anyway.
Finally, you seem to have a poor understanding of how social media works, another invention with immense power, which amplifies the spread of extremist opinions. In the cyberspace of social networks, being reasonable means going unnoticed, be silenced and therefore having no real power of influence.
The Federal Contracting system could have been quietly used to redirect which contractor gets what pot of money on a “Needs of the Government” basis. It’s not difficult, it is not routine, but not difficult. The power is already built into the procurement system.
Hegseth’s performative Theater Savant is in another realm, as is Amodei’s very public response, and Altman’s machinations.
Hegseth provides Trump / Miller / Bannon et al. with the in-your-face Dumb Guy persona to keep the Anti-Woke farce alive. He’s an utter moron, the Pentagon knows it, everyone’s pretty sure that he might also know it, though Pete must “soldier” on to impress his audience of one lest he be cut loose. There were plans for him to see the underside of the bus in late January, but hey… opportunistic war time.
Amodei’s moral stance is likewise staged. The protections that he references are already built into the legal system , and LLMs don’t control autonomous targeting systems. That’s not how one gets to Death by Robot.
Anthropic has portrayed itself as being the moral alternative to Clammy Sammy’s AI finance carousel. In fact, OpenAI and Anthropic are essentially branches of the same sham run by a pair of monumentally uncharming hucksters. Individuals I know who’ve interfaced with both find Altman incredibly unlikable and Amodei exponentially more distasteful, so once it really becomes clear that this is all a pyramid scheme… neither of their personalities are going to keep the public funding machine running.
OpenAI’s debt is astronomical. Not just massive but defies the basic rules of math as to how any of this works out in positive returns. Ever…
One of the players that is involved in the circular debt-for-equity farce is Ellison’s Oracle. Oracle is staring down the barrel of multiple lawsuits now that its bondholders have gone all torches and pitchforks after figuring out what going on. Hell hath no fury like Tech investors scorned.
Which brings us back to Clammy Sammy…
If OpenAI goes under, the whole finance scam falls apart and likely collapses that large part of the US market based on the AI oversell.
It has been deemed “unsightly” by the Trump folks to initiate a bailout for Sam and the AI train (they’re saving that maneuver for Tether).
Thus, Sam and OpenAI will be eligible for that same sweet pot of dollars via the Feds once they are on a “Cost Plus” contract vehicle in which the government pays for whatever the contractor’s expenses are and the “Plus” of whatever fee the contractor and the Feds work out to be reasonable. In the old world, the fee was negotiated, in today’s world where the OMB and other agencies have simply become grift machines… who TF knows what will happen?
As proven myriad times, it’s easy to redirect funds between Federal money pots. It’s not always legal, but that hasn’t stopped the Trumpers yet. So…
So, in summary, Hegseth’s chest thumping pronouncements were just theater to shift contracts from Anthropic to OpenAI. Why? Because Weird Sam is out of money and the next funding round was looking really bad in terms of investor confidence, and Larry Ellison needs guarantees of even more Fed cash to pacify his enraged bondholders.
Because, as has always been known by anyone who has used a government system powered by Oracle…
Oracle doesn’t have clients, they have hostages.
And Anthropic continues to pretend to be the measured / moral alternative to OpenAI (they aren’t) to attract the next X Billion (30? 50? Who knows?) that they will need shortly, because their fee math is as sketchy as OpenAI’s. Anyone finding Claude useful enough to pay twenty to fifty times as much for it to keep the compute going? Just curious.
I feel as if some sort of government/AI company showdown was always going to happen because governments usually don’t like it when private entities and individuals control powerful, humanity altering technology. In a normal administration, this would
probably involve a long drawn out process of reigning in and regulating tech companies (or not). Instead, they went all in for a man they believed would de regulate their industry AND leave them alone to do as they pleased.
“Simply put: Claude can make mistakes, and when Claude is used in applications where a mistake has the potential to irrevocably change existing human lives, a human or company should be accountable for those mistakes.”…no truer words have been spoken than these. It irritates me to no end those who claim that the liability would be too large, and subject to regulations in support of this, it would stunt the development of these “life changing” technologies. To that I say, “good!”, perhaps it would force them to look deeper and work harder to make sure these AI technologies are well understood, predictable, and controllable. Instead we get their claims that it could doom society, but they’ll keep moving things right ahead, as though they have no choice but to proceed irrespective of the potential dangers that they themselves acknowledge.
As for this predictable outcome from Hegseth, that was telegraphed with Trump’s EO and his appointment of tech utopian David Sacks as Czar of tech pipe dreams.
Brilliant.
Dear Dr Casey Mock, you begin by pitting the extremists back to back, the techno-optimists (accelerationists) and the Cassandres (doomers), to which I belong, along with my former PhD supervisor in Montréal. Admittedly, that's easy!
Then, in the rest of his post, you attack the Cassandres and proposes a more reasonable and balanced approach, arguing that AI is a technology like any other, useful but nothing more. You essentially criticize the Cassandres for highlighting the power of AI technology, which attracts power-hungry individuals and "dark triad" types (egocentric, manipulative, psychopathic) like the Secretary of War.
On this point, you're probably wrong. Trivializing AI is like comparing the invention of dynamite to that of the atomic bomb. There are orders of magnitude of difference and potential negative impacts. The same is true for those comparing the use of calculators and generative AI in education. Not to mention the effect of intellectual laziness on our brains.
AI has the potential to change the world, but it is up to us, Humans, to democratically decide which direction to take in the best interest of all, not just a handful of technobillionaires.
Even if the Cassandres "would try to outsmart" by trivializing AI, die-hard techno-optimists like Sam Altman, who promise nothing less than "eternal life" and an "infinite cornucopia" (sic, in a finite world), will have convinced the power-hungry ones anyway.
Finally, you seem to have a poor understanding of how social media works, another invention with immense power, which amplifies the spread of extremist opinions. In the cyberspace of social networks, being reasonable means going unnoticed, be silenced and therefore having no real power of influence.
Claude, in the flesh, COULOMBE
Well…
The Federal Contracting system could have been quietly used to redirect which contractor gets what pot of money on a “Needs of the Government” basis. It’s not difficult, it is not routine, but not difficult. The power is already built into the procurement system.
Hegseth’s performative Theater Savant is in another realm, as is Amodei’s very public response, and Altman’s machinations.
Hegseth provides Trump / Miller / Bannon et al. with the in-your-face Dumb Guy persona to keep the Anti-Woke farce alive. He’s an utter moron, the Pentagon knows it, everyone’s pretty sure that he might also know it, though Pete must “soldier” on to impress his audience of one lest he be cut loose. There were plans for him to see the underside of the bus in late January, but hey… opportunistic war time.
Amodei’s moral stance is likewise staged. The protections that he references are already built into the legal system , and LLMs don’t control autonomous targeting systems. That’s not how one gets to Death by Robot.
Anthropic has portrayed itself as being the moral alternative to Clammy Sammy’s AI finance carousel. In fact, OpenAI and Anthropic are essentially branches of the same sham run by a pair of monumentally uncharming hucksters. Individuals I know who’ve interfaced with both find Altman incredibly unlikable and Amodei exponentially more distasteful, so once it really becomes clear that this is all a pyramid scheme… neither of their personalities are going to keep the public funding machine running.
OpenAI’s debt is astronomical. Not just massive but defies the basic rules of math as to how any of this works out in positive returns. Ever…
One of the players that is involved in the circular debt-for-equity farce is Ellison’s Oracle. Oracle is staring down the barrel of multiple lawsuits now that its bondholders have gone all torches and pitchforks after figuring out what going on. Hell hath no fury like Tech investors scorned.
Which brings us back to Clammy Sammy…
If OpenAI goes under, the whole finance scam falls apart and likely collapses that large part of the US market based on the AI oversell.
It has been deemed “unsightly” by the Trump folks to initiate a bailout for Sam and the AI train (they’re saving that maneuver for Tether).
Thus, Sam and OpenAI will be eligible for that same sweet pot of dollars via the Feds once they are on a “Cost Plus” contract vehicle in which the government pays for whatever the contractor’s expenses are and the “Plus” of whatever fee the contractor and the Feds work out to be reasonable. In the old world, the fee was negotiated, in today’s world where the OMB and other agencies have simply become grift machines… who TF knows what will happen?
As proven myriad times, it’s easy to redirect funds between Federal money pots. It’s not always legal, but that hasn’t stopped the Trumpers yet. So…
So, in summary, Hegseth’s chest thumping pronouncements were just theater to shift contracts from Anthropic to OpenAI. Why? Because Weird Sam is out of money and the next funding round was looking really bad in terms of investor confidence, and Larry Ellison needs guarantees of even more Fed cash to pacify his enraged bondholders.
Because, as has always been known by anyone who has used a government system powered by Oracle…
Oracle doesn’t have clients, they have hostages.
And Anthropic continues to pretend to be the measured / moral alternative to OpenAI (they aren’t) to attract the next X Billion (30? 50? Who knows?) that they will need shortly, because their fee math is as sketchy as OpenAI’s. Anyone finding Claude useful enough to pay twenty to fifty times as much for it to keep the compute going? Just curious.
Thanks for your efforts on the piece.
All the best.
For Hegseth, Doom™️ is a feature and not a bug.