AI meets the Department Of Government Efficiency
I have a reflexive skepticism about fads. While AI isn’t exactly a fad, the relentless promotion by the media of an AI future certainly looks faddish to me.
That said, there’s no question that AI is making inroads into data analysis. Watson can already read radiological images more accurately than trained radiologists. AI is transforming warehouse management.
Over at her blog “Eating Policy,” Jennifer Pahlka has a post on the role AI is likely to play in the execution of the Musk/Ramaswami DOGE project to cut $2 trillion from the federal budget.
“The reality is that in many domains, the regulatory and spending complexity is such that it’s very hard for anyone to know what’s going on. You might think it’s Congress’s job to understand how the laws they’ve written have been operationalized, but that’s one of their chief complaints — that they don’t really understand what happens within the agency and they don’t always think it’s consistent with their intentions. And the agencies themselves are dealing with the accretive nature of what comes down from Congress — new laws naturally reference and amend old laws, creating one confusing web of language. Then there’s the web of the regulations previous staff have written, not to mention the policies, forms, and processes that have been born from those regulations that seem to carry the weight of the law but are really somewhat arbitrary expressions of one way they could be operationalized. It becomes hard to sort out what cannot be changed without an act of Congress from what would need a new rule (and therefore a rulemaking process) from what could entirely legally be modified if only Bob over in compliance would stop threatening to call the Inspector General.”
Elon is probably already on this, training AI to parse what is regulatory law from what is merely current practice:
“Imagine DOGE walking into agencies on January 21st and not having to say [“show me where it says we can’t do this”] four times a day. If they’re building good AI models (and, let’s hope, testing them), they’re not going to ask that at all. They’re going to know what’s legal and what’s not, or at least think they’ll know. (It’s all always open to interpretation.) Right there, the wall I talked about on Friday is immediately pierced. It’s not so much an information asymmetry we’ll be looking at, but an asymmetry of understanding, and of confidence (merited or not) in their ability to act, and act fast.”
Congress is not ready to rebut this challenge or to respond in a timely fashion by creating law to back its intentions.
“Most commentary on DOGE has pitted it against the agencies it has vowed to drastically cut, but Congress is going to want to have a say in what they propose to do — and of course, the executive branch is legally bound to spend what Congress has appropriated, though the Trump administration has signaled an interest in challenging (or ignoring) the Impoundment Control Act of 1974 which, according to a statement from the Democrats in the House Committee on Budget, “established procedures to prevent the President and other government officials from unilaterally substituting their own funding decisions for those of the Congress.” In other words, if the budget passed by Congress says we’re spending this, we’re spending this, DOGE or Trump be damned.”
Of course, the GOP Congress can’t even pass a continuing resolution these days. This ain’t your father’s budget process, peeps.
AI, DOGE and the federal budget
That said, there’s no question that AI is making inroads into data analysis. Watson can already read radiological images more accurately than trained radiologists. AI is transforming warehouse management.
Over at her blog “Eating Policy,” Jennifer Pahlka has a post on the role AI is likely to play in the execution of the Musk/Ramaswami DOGE project to cut $2 trillion from the federal budget.
“The reality is that in many domains, the regulatory and spending complexity is such that it’s very hard for anyone to know what’s going on. You might think it’s Congress’s job to understand how the laws they’ve written have been operationalized, but that’s one of their chief complaints — that they don’t really understand what happens within the agency and they don’t always think it’s consistent with their intentions. And the agencies themselves are dealing with the accretive nature of what comes down from Congress — new laws naturally reference and amend old laws, creating one confusing web of language. Then there’s the web of the regulations previous staff have written, not to mention the policies, forms, and processes that have been born from those regulations that seem to carry the weight of the law but are really somewhat arbitrary expressions of one way they could be operationalized. It becomes hard to sort out what cannot be changed without an act of Congress from what would need a new rule (and therefore a rulemaking process) from what could entirely legally be modified if only Bob over in compliance would stop threatening to call the Inspector General.”
Elon is probably already on this, training AI to parse what is regulatory law from what is merely current practice:
“Imagine DOGE walking into agencies on January 21st and not having to say [“show me where it says we can’t do this”] four times a day. If they’re building good AI models (and, let’s hope, testing them), they’re not going to ask that at all. They’re going to know what’s legal and what’s not, or at least think they’ll know. (It’s all always open to interpretation.) Right there, the wall I talked about on Friday is immediately pierced. It’s not so much an information asymmetry we’ll be looking at, but an asymmetry of understanding, and of confidence (merited or not) in their ability to act, and act fast.”
Congress is not ready to rebut this challenge or to respond in a timely fashion by creating law to back its intentions.
“Most commentary on DOGE has pitted it against the agencies it has vowed to drastically cut, but Congress is going to want to have a say in what they propose to do — and of course, the executive branch is legally bound to spend what Congress has appropriated, though the Trump administration has signaled an interest in challenging (or ignoring) the Impoundment Control Act of 1974 which, according to a statement from the Democrats in the House Committee on Budget, “established procedures to prevent the President and other government officials from unilaterally substituting their own funding decisions for those of the Congress.” In other words, if the budget passed by Congress says we’re spending this, we’re spending this, DOGE or Trump be damned.”
Of course, the GOP Congress can’t even pass a continuing resolution these days. This ain’t your father’s budget process, peeps.
AI, DOGE and the federal budget
You have a lot more faith in AI than I. Do you really think it can follow the necessary legal and real world reasoning when it is much more likely to simply hallucinate and come up with something that sounds reasonable but is likely incorrect? Odds are the AI system will simply cite some legal sounding gibberish rather than actually citing the law as it is written.
@Kaleberg,
I hope you’re right and that Musk is diving down an AI rabbit hole. It has everything to do with training AI. And in the short term, it just has to be better than Congress, which is a pretty safe bet.
RJ:
You would have to give me the whole thing and I would take it off site to fix it. There is too much going on here for me to sort out.
i’ll just paste the same thing over without any html code, Bill, then you can delete that mess…
after i hadn’t posted a link here in a while, Bill, i forgot that i could just paste it like that…most websites require the html coding to include links in comments, so that’s what i tried to do…so now you can just delete the above, and these two comments of mine about it, and be done with it…
sorry for all the trouble…
rjs:
It is ok. Usually, I can fix simple to medium difficult things. This one was a little more complex. I would have had to take it offline to do so. Sorry for not being able to do so. Repost was the best solution. I deleted the old version.
i had been using Microsoft’s AI, called copilot, which they installed on my new Windows 11 machine, to research the latest Covid mutations while writing the brief updates on the latest Covid trends that i email out each week…up until a couple months ago, i thought the info i was getting from that AI gadget was fairly dependable, but in the last two exchanges i had with it, it made serious mistakes….first it told me that LB.1.3.1 is a descendant of KP.2, which it is not…it’s a direct descendant of JN.1, whereas KP.2 evolved from JN.1.11.1…a week later, it gave me info about one mutant Covid variant (XEC) when i asked about a different one (MC.1) – three times…it also says MC.1 is a recombinant strain – it is not – then it told me the MC.1 variant has unique mutations, but it couldn’t tell me what those mutations are…one must know what they are to call them unique..
a little data on AI data centers:
a couple thoughts…if energy demand from data centers triples in four years, utilities and the grid won’t keep up…some may build dedicated nuclear plants to run their data centers, but a lot are going to compete for already available resources, which is why old the mothballed coal and nuclear power plants are coming back…trying to run those old nuke plants that were designed to last just 40 years for 80 years is an accident waiting to happen…
the AI spending numbers are stunning…those four companies aren’t spending over $100 billion to enhance my search capabilities; they must expect a profitable return from something else…also note that a big chunk of that spending is still for teaching AI everything we know…so this thing is still in its infancy…& no one has any idea where it’s going; they just don’t want the other guy to get there first…