Personal Finance Topics / Macroeconomic Trends and Risks
No. of Recommendations: 7
Minister of Defense of Ukraine: Ukraine opens battlefield data to train AI models for autonomous systems — first in the world. Millions of annotated combat frames from thousands of missions are now available for partners to train AI. This creates a new win-win cooperation model: partners refine their technologies with unique data, while Ukraine scales frontline capabilities.
Anthropic has potentially committed commercial suicide to not be used for this purpose in the US because they felt they could not guaranty their software would behave itself, but OpenAI seems to have chosen the side with the money bags.
I had dinner last night with a former US military procurement bigwig and she was adamant that the position taken by the DoD was the correct one and it was unconscionable for a vendor to constrain the use of their product (even from mass surveillance of the US population).
Warfare is about to get real interesting.
Jeff
No. of Recommendations: 0
Warfare is about to get real interesting.
Especially if it targets war-favoring politicians in the US.
No. of Recommendations: 12
" the position taken by the DoD was the correct one and it was unconscionable for a vendor to constrain the use of their product (even from mass surveillance of the US population)."
She is wrong. A US with those capabilities is no longer a US. Anthropic has taken a proper moral stand. If only other companies were so nobly inclined (not just AI companies).
Dangerous times.
No. of Recommendations: 4
> She is wrong. A US with those capabilities is no longer a US. Anthropic has taken a proper moral stand. If only other companies were so nobly inclined (not just AI companies).
> Dangerous times.
I try to post when I have information to add, or a novel viewpoint. Sometimes I fail, but I try.
This is an intentional exception, though.
I wholeheartedly agree with you and felt it was important to speak up. Using these information doomsday weapons as glibly as the current leadership is will lead to leading researchers, both foreign and domestic, leaving the US to research and work elsewhere. China, India, and Europe are hiring, paying, and are far different than the average mass media consumer in the US has been led to believe. This would be a substantially larger, national scale example of how PLTR has recruitment issues even with its notoriously high compensation, because of the stink of the name on resume.
The DoD's stance was morally and - almost worse - strategically wrong. Stupid, self-defeating, and *uncharacteristically and embarrassingly public* behavior. Clearly DoD leadership has changed in quality and character.
What I can offer is that OpenAI in some areas seems better connected to Republican-corporate power brokers. This seemed contrived and a giveaway to political allies. One can follow model performance closed and OpenAI has fallen beyond almost all its rivals. The scaled back data center plans with Oracle are just one example. If you know the name Sam Altman and followed the OpenAI Foundation PBC politics over the last 5 or so years you'll know about as much as you need to know about the company. They have hemorrhaged top research talent since then, as those people can write their own offers at any frontier research lab or startup and can vote on ethics with their feet.
No. of Recommendations: 1
No. of Recommendations: 5
> Seems?
Those are more public examples. That is at most half of what Brockman has donated. That is direct support, and symbolic. The real support is via lobbying, which is much, much larger and connects very directly to policy decision makers.
SamA's donation is just performative ring- and ass-kissing.
Think more about the lobbying, the federal contracts, and the players in Stargate.
Anyway, this is METAR: connecting it to macro trends - consider AI and AI related companies very volatile in the near term because it is so innately connected to funding in an industry operating at loss and trying to make it up on scale. Keep an eye on
* levels of token subsidy by foundational model platforms
* Ali Cloud research and Qwen, specifically, among international competition
* data center write downs and cancelled infrastructure products
AI is real. AGI may yet be real, and distinction may not be clear. The degree to which it is profitable in the long term remains to be seen. The massive compute investment will suffer real depreciation faster than other hardware. GPUs and NVMe under heavy usage will have a duty cycle less than 5 years. NVDA, STX, WDC will do well while the funding and corporate mandates last. MU may have mixed headwinds and tailwinds. They are vulnerable to growth falling below projections.
A heavy handed and heavily partisan approach to the US AI industry is going to be extremely harmful to a field that is in hot international competition for a small pool of elite research talent, is losing the trust of a population feeling layoffs and increased power and water bills, and is getting unpopular for moronically clumsy usage (Flock, Ring, ICE's Palantir apps, generated targeting that killed the 200 little girls in Minab, et alia).
If the US falls behind in this area - and it is at real risk of rapidly losing its early lead - it may be a difficult gap to close.
No. of Recommendations: 1
Ormont, with due respect,
Not everything is a commercial issue. Humanity isn't. Life on earth isn't. Your US military procurement bigwig is talking her book. So, what do you expect???
Warfare is about to get interesting [real] because there are a lot of unethical and evil people in the world whose only interest is their own power. And now they are trying to corner the information universe so they can program and control the means and the enforcement against any human they target, not just another country.
Self-interest at that level is not interesting, its evil and must be stopped, whoever they are.
Education is a good place to start. AI is not intelligence... It's uber programming and algorithms for mass control by the few--who are human, flawed and self-interested. Let em develop some robo-cops, why don't we??? And give em every detail about our lives.
Hopefully, none of us inadvertently think any bad thoughts, huh?
No. of Recommendations: 2
No. of Recommendations: 6
Thanks for that. I haven't seen that one.
Thinking we already have our AI viruses, in human form. Peter Thiel, Trump, Vance, Musk, Zuckerberg, just to name a few. Whatever they touch is trouble for us.
