It is about certain regimes nearing their end and folks converting assets into something fungible they can use and enjoy while exiled in Geneva, Dubai, Phnom Penh, etc.
As just one example (of many) of why its not about the USD - most global debt is dollar denominated and settled in dollars. Even if they don't reside or transact in the US, most large financial transactions settle (or are hedged) in USD. Again, just one example.
Also, understand export economies like China cannot avoid dollar settlement for goods exported to the US. They can settle in USD and covert to another asset, but only as a secondary step.
I could go on about this, but Carnegie Endowment Prof Michael Pettis explains this and more much better than I can.
I expected to see OpenAI, Google, Anthropic, etc. provide desktop applications with integrated local utility models and sandboxed MCP functionality to reduce unnecessary token and task flow, and I still expect this to occur at some point.
The biggest long-term risk to the AI giant's profitability will be increasingly capable desktop GPU and CPU capability combined with improving performance by local models.
From experience it seems like preempting context scoping and routing decisions to smaller models just results in those models making bad judgements at a very high speed.
Whenever I experiment with agent frameworks that spawn subagents with scoped subtasks and restricted context, things go off the rails very quickly. A subagent with reduced context makes poorer choices and hallucinates assumptions about the greater codebase, and very often lacks a basic sense the point of the work. This lack of situational awareness is where you are most likely to encounter js scripts suddenly appearing in your Python repo.
I don’t know if there is a “fix” for this or if I even want one. Perhaps the solution, in the limit, actually will be to just make the big-smart models faster and faster, so they can chew on the biggest and most comprehensive context possible, and use those exclusively.
eta: The big models have gotten better and better at longer-running tasks because they are less likely to make a stupid mistake that derails the work at any given moment. More nines of reliability, etc. By introducing dumber models into this workflow, and restricting the context that you feed to the big models, you are pushing things back in the wrong direction.
Good. I want my tax dollars allocated to penetrating every and any system my country's adversaries may use to undermine our interests or threaten our people. And, I want maximum penalties, civil and criminal, for any person or company who misuses these systems for personal or political gain. Also, I'd like to see mandatory statutory civil damages for any vendor creating and/or selling/providing these systems who does so in a negligent or malicious manner, same as we provide for other high risk products and services.
Well, you're definitely not going to get the latter two, and the only guarantee about the first one is that they will definitely be used against enemies of the state.
Whether there's any overlap between them and enemies of the people will heavily depend on the latter's ability to steer towards good governance. The track record for the past few decades hasn't been great.
Then they are not the best. Most users aren't prompt engineers and grew up expecting to enter search terms into Google and get a result. If its the case OpenAI or Anthropic are best able to interpret user intent there's a good argument to be made they are the best.
If model trusts the users, and if user is dumb model will "weigh" user's input much higher and end up with flawed code.
If the model is more independent, it will find the right solution. If just want a dumb model which says yes to everything, and follows you when u are not at smart enough then you'll never end up with good solution if not by luck.
Have a look at industrial accident data globally, considering underreporting in developing nations.
- Same, environmental accidents.
- Same, WMD proliferation, including chem, bio and nuclear.
- Same, malicious cyber.
Now, ask yourself if we have enough problems aligning & regulating AI at the moment?
Are we sure that in the name of laudable egalitarian ideals that we are prepared for the second and third order effects of broad global accessibility to AI, including frontier models?
Notes: I'm stuck with this and wish I could leave, but the combo of security and attachments leaves me with few options. Notes is buggy, lacks interoperability, has crap AI integration, etc. BUT... the real pain is Apple Messages. The inability create folders, rules, etc. is insanity. At a time when most comms are instant why would you not offer users the same level of organization they enjoy with email, or with other IM platforms?
Free market is better when its truly free. Free from:
- Predatory export subsidies.
- Protective non-monetary barriers to import/export.
- Theft of intellectual property.
- Monopolization of dual-use and essential manufactured products.
- Two sides that are not posturing for war.
- One side that is not disregarding international law.
In the case of AI, the scale of CCP military command & control, and intelligence collection is vast due the size of the PLA, PLAN, MSS, etc. Denying them AI to fuse and coordinate may be a lost cause ultimately, but time has a value all its own and that's what the US is after as it - and its allies - reconfigure to defend the SCS, India & the Pacific in general.
Sabine highlights the problem with scientific funding in this video and it should be required watching before posting on this thread. Reform is needed. Some good will be tossed with a lot of bad. Its a cycle, a pendulum, and it will eventually tip to excess again sometime in the future. For now... fixing what is broken ought to be the priority.
I got to the point where he says the email she made a video about is probably her own making and stopped. It is a >1.5h response to a 10-minute video, and at minute 7 (of 1.5h) he proceeds to basically call her a liar in a fundamental way without having any strong evidence for it. Mind you the first 7 minutes were spent claiming she put forall quantifiers where she just implied strong prevalence and telling how much better the guy is for the society than her. From these things I would say you'd waste your time watching this unless you want to practice fishing for fallacies. The 3 I mentioned so far are: the leaky bucket fallacy (weak evidence for liar claim), straw man (arbitrary adding forall), and ad hominem (attack character instead of presenting argument).
"The field" in her case is "particle physics". And she's been making a very good case against the non-science being done in that field. Unfortunately, like physicists tend to do, for some reason, she's branched out into criticizing "not her field" as well, sometimes even non-science topics, to far worse effect. She's become an excellent example of audience capture, a loss to us all (and a loss to credibility she earned within particle physics).
I think there may be a language issue here; to use her own words as best as I can remember them, excusing her bluntness under "perhaps I'm just German" — a messy kichen here in Germany would be described with the word "Chaos", and a mistake that a Brit would call "dropping the ball" would be described as "eine totale Shitshow".
But that means I don't put too much credence to her summary of climate science or trans stuff: when it's the topic of inclusivity attempts, she's got the direct personal experience to play the "here's how well intentioned policies backfire" card; when it's the internal politics within science, honestly that reminds me a lot of software development's cycle of which language, framework, design pattern, and organisational orientation pattern (objects, composition, functional, etc.) is a code smell or the smell of coffee that one should wake up to, so it rings true even if I can't verify it.
Yes, we create service accounts on the source platforms which can then be used to do an OAuth or key based integration. What would you like to do specifically?
It is about certain regimes nearing their end and folks converting assets into something fungible they can use and enjoy while exiled in Geneva, Dubai, Phnom Penh, etc.
As just one example (of many) of why its not about the USD - most global debt is dollar denominated and settled in dollars. Even if they don't reside or transact in the US, most large financial transactions settle (or are hedged) in USD. Again, just one example.
Also, understand export economies like China cannot avoid dollar settlement for goods exported to the US. They can settle in USD and covert to another asset, but only as a secondary step.
I could go on about this, but Carnegie Endowment Prof Michael Pettis explains this and more much better than I can.
reply