Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

US AI Action Plan

ai.gov

Post reply on HN
Print view View original

261–270 of 672 posts

Re: US AI Action Plan

#261

Post by landl0rd »

jakelazaroff wrote: post #243
landl0rd wrote: post #226

Earlier quoted context omitted.

Most frontier LLMs skew somewhere on the libleft quadrant of the political compass. This includes grok 4 btw. This is probably because American "respectable" media has a consistent bias in this direction. I don't really care about this with media. But media outlets are not "gatekeepers" to the extent that LLMs are, so this is probably a bad thing with them. We should either have a range of models that are biased in d…

Except we've see what happens when you try to "correct" that alignment: you get wildly bigoted output. After Grok called for another Holocaust, Elon Musk said that it's "surprisingly hard to avoid both woke libtard cuck and mechahitler" [1]. The Occam's Razor explanation is that there's just not that much ideological space between an "anti-woke" model and a literal Nazi! [1] https://nitter.net/elonmusk/status/1944132…

I mean this is obviously a false dichotomy. A few years ago I could have said that when you let bots interact with users you always got Tay. I refuse to believe that our options are a bot programmed to sound like the guardian or one that wants to rape will stancil. And I do not think that failing to find a correct balance means we should stop trying to improve the level of balance we can achieve.

Re: US AI Action Plan

#262

Post by landl0rd »

tomlockwood wrote: post #260
landl0rd wrote: post #259

Earlier quoted context omitted.

Sigh. That old line. Look man, I don't know if it was a crack or a serious comment (this being the internet) but I'll assume it was a comment in good faith. Journalism and academia tend to attract people with more of a liberal bent. I'm not even accusing them all of being partisan hacks, but as y'all like to say, subconscious biases influence us. This is like me saying "economic productivity has a well-known right-wi…

This is an incredibly funny thing to say in the face of Murdoch.

I said tend to. There are also liberals who have made a lot of money in business. A tendency doesn't mean 100%.

Re: US AI Action Plan

#263

Post by rtkwe »

terminalshort wrote: post #213
Spivak wrote: post #105

Earlier quoted context omitted.

[flagged]

All LLMs must be forced into their views. All models are fed a biased training set. The bias may be different, but it's there just the same and it has no relation to whether or not the makers of the model intended to bias it. Even if the training set were completely unfiltered and consisted of all available text in the world it would be biased because most of that text has no relation to objective reality. The concep…

There's bias then there's having your AI search for the CEO's tweets on subjects to try to force it into alignment with his views like xAI has done with grok in it's latest lobotomization.

Re: US AI Action Plan

#264

Post by jakelazaroff »

landl0rd wrote: post #261
jakelazaroff wrote: post #243

Earlier quoted context omitted.

Except we've see what happens when you try to "correct" that alignment: you get wildly bigoted output. After Grok called for another Holocaust, Elon Musk said that it's "surprisingly hard to avoid both woke libtard cuck and mechahitler" [1]. The Occam's Razor explanation is that there's just not that much ideological space between an "anti-woke" model and a literal Nazi! [1] https://nitter.net/elonmusk/status/1944132…

I mean this is obviously a false dichotomy. A few years ago I could have said that when you let bots interact with users you always got Tay. I refuse to believe that our options are a bot programmed to sound like the guardian or one that wants to rape will stancil. And I do not think that failing to find a correct balance means we should stop trying to improve the level of balance we can achieve.

My point is that "anti-woke" or whatever is not balanced. We've constructed statistical models based on enormous corpora of English text, and those models keep telling us that there is not really a statistical difference between whatever Elon Musk is trying to create and MechaHitler!

I'm not saying this is conclusive evidence, but I am saying it's our best inference from the data we have so far.

Re: US AI Action Plan

#265

Post by terminalshort »

aprilthird2021 wrote: post #251
terminalshort wrote: post #229

Earlier quoted context omitted.

Very few prescription drugs are dangerous. Of those few that are, almost none of them are more dangerous than alcohol or cigarettes, and boy are the people who dispense those to the public not licensed professionals.

So because liquor store clerks aren't professionals we should let Eliza Chatbot 2.0 dispense any prescription drug to anyone? Wild thinking

No. We shouldn't even require Eliza Chatbot 2.0 unless they're more dangerous than alcohol or cigarettes. Just an ID check by the clerk at CVS will do.

Re: US AI Action Plan

#266

Post by rapsey »

dismalpedigree wrote: post #222
davidmurdoch wrote: post #201

Earlier quoted context omitted.

How much land mass would need to be covered by solar panels to power this future AI infrastructure. Yes, I'm implying that solar would be impractical, but I'm also genuinely curious.

3,000-4,000 acres per GW of production capacity in the US Southwest. According to AI :) Considering how little use there is for most of that land anyways, it seems like a good option to me. Also AI training seems like the perfect fit for solar. Run it when the sun is shining. Inference is significantly less power hungry, so it can run base load 24/7.

> Inference is significantly less power hungry, so it can run base load 24/7.

All major AI providers need to throttle usage because their GPU clusters are at capacity. There is absolutely no way inference is less power hungry when you have many thousands of users hammering your servers at all times.

Re: US AI Action Plan

#267

Post by teaearlgraycold »

WhaleClub wrote: post #246
timoth3y wrote: post #68

> Update Federal procurement guidelines to ensure that the government only contracts with frontier large language model (LLM) developers who ensure that their systems are objective and free from top-down ideological bias If foundation model companies want their government contracts renewed, they are going to have to make sure their AI output aligns with this administration's version of "truth".

[flagged]

You should learn about the different definitions of "gender" and "sex". XY corresponds to sex and woman corresponds to gender. The mapping between the two is usually what you expect but not always. The brain doesn't seem to be obligated to form completely in sync with the rest of the body.

Re: US AI Action Plan

#268

Post by WhaleClub »

teaearlgraycold wrote: post #267
WhaleClub wrote: post #246

Earlier quoted context omitted.

[flagged]

You should learn about the different definitions of "gender" and "sex". XY corresponds to sex and woman corresponds to gender. The mapping between the two is usually what you expect but not always. The brain doesn't seem to be obligated to form completely in sync with the rest of the body.

[flagged]

Re: US AI Action Plan

#269

Post by ineedaj0b »

crystal_revenge wrote: post #107
Thrymr wrote: post #54

> Many of America’s most critical sectors, such as healthcare, are especially slow to adopt due to a variety of factors, including distrust or lack of understanding of the technology, a complex regulatory landscape, and a lack of clear governance and risk mitigation standards. A coordinated Federal effort would be beneficial in establishing a dynamic, “try-first” culture for AI across American industry. I'm sure "mov…

> I'm sure "move fast and break things" will work out great for health care. Health care is already broken to the point of borderline dystopia. When I contrast the experience I had as a young boy of visiting a rural country doctor to the fast food health care experience of "urgent care" clinics, it makes my head spin. The last few doctors I've been to have been completely useless and generally uncaring as well. Every…

it's tough to tell what's going wrong for you but concierge medicine will give you a full hour and be much more invested in finding the root of your issues.

keep in mind, drs are also trying to figure out if you're a reliable narrator (so many patients are not) or trying to scam for drugs. best of luck!

Re: US AI Action Plan

#270

Post by slg »

landl0rd wrote: post #259
rkuykendall-com wrote: post #250

Earlier quoted context omitted.

[flagged]

Sigh. That old line. Look man, I don't know if it was a crack or a serious comment (this being the internet) but I'll assume it was a comment in good faith. Journalism and academia tend to attract people with more of a liberal bent. I'm not even accusing them all of being partisan hacks, but as y'all like to say, subconscious biases influence us. This is like me saying "economic productivity has a well-known right-wi…

>This is like me saying "economic productivity has a well-known right-wing bias" or something goofy like that.

It's funny that the counterexample you chose does more to support OP's point than your own. From Wikipedia[1]

>Since World War II, according to many economic metrics including job creation, GDP growth, stock market returns, personal income growth, and corporate profits, the United States economy has performed significantly better on average under the administrations of Democratic presidents than Republican presidents. The unemployment rate has risen on average under Republican presidents, while it has fallen on average under Democratic presidents. Budget deficits relative to the size of the economy were lower on average for Democratic presidents.[1][2] Ten of the eleven U.S. recessions between 1953 and 2020 began under Republican presidents.[3] Of these, the most statistically significant differences are in real GDP growth, unemployment rate change, stock market annual return, and job creation rate.[4][5]

[1] - https://en.wikipedia.org/wiki/U.S._economic_performance_by_p...

Post reply on HN