[ben_w]

> I there's anyone I don't trust with AI, it's the worlds #1 company in spying on people, in collection of Pii, in tracking, and many many many times caught literally lying about it.

If you gave me this description in isolation, I would assume you were talking about Meta, who are not on your shortlist despite also having AI models.

However, this is beside the point. Note my careful phrasing: I was not disagreeing with you asking "Who are the good actors within the space?" (which I understood as a rhetorical way of saying "none of them"), I was only rank-ordering them.

To me, it appears Grok isn't even putting in effort to try to look trustworthy. China is trying, though there is suspicion. OpenAI likes to say the right things, though (while I cannot see it myself with regards specifically to his work at OpenAI*) most commentary about Altman regards him as a wrong'un, the phrase "King of the cannibals" comes to mind, and the CEO is limiting trust in the company**.

Comparing Anthropic and Google: Anthropic looks to me like they're actually trying to do the right thing even when it's expensive and painful, but some hate them just for being in AI at all. Google appears to still have a competent PR department, though I don't trust them myself. So, I'd rank Anthropic then Google, but I think there's many who would rank Google then Anthropic.

* But I do put about 40% odds on his sister's allegations being true.

** My guess here is that we all saw Zuckerberg being presented as a heroic underdog and how that turned out, and naturally people don't want a repeat of this.

[b112]

If you gave me this description in isolation, I would assume you were talking about Meta, who are not on your shortlist despite also having AI models.

You've mistaken me for the guy you originally replied to. It happens, no worries, but I'm not that guy. I'm just replying to the list I saw you state.

Between Meta and Google? Oh sure, Meta collects info for sure. But Meta doesn't lie, abuse, and misuse the Google Play framework to collect data, it doesn't use the same via push notifications to collect enormous amounts of data, nor does Meta host endless fonts and libraries and DNS infra, solely for the purposes of tracking what hosts/users do.

Beyond that, whilst Meta and Google both work to embed stuff in webpages for additional tracking, Google sign-on is another massive tracking infra. And of course, gmail endlessly scanned and data snarfed, including for AI training, which is immensely beyond Meta's capabilities. And firebase?! Oh geez.

"Hi, I'm Google! I'm going to provide this neat thing called firebase, so every single high security application, banking to identity verification, will include our tracking software in it for device attestation and push notifications! But of course, we'll never use that to track anyone! Honest!"

Compared to Google's data collection infrastructure, Meta is a tiny, miniscule gnat in comparison. And yes, I know how bad Meta is. If anything, I think Meta is at least quite honest about much of what data is collected. Zuck isn't a saint, but it's not like he has entire operating systems sneakily stealing user location data and then lying about doing so. Then saying they fixed that, and being caught a few more times, each with "Oh, so sorry, we'll fix that now. Really". That's Google, and that attitude is prevalent in their entire org.

To use Google anything, is to feed the most powerful, wide spread, incredible data collection machine the world has ever seen. It dwarfs everyone and everything else, factorially. I bet if you took all the data that the NSA and every other single spy agency on the planet has on its citizenry, they'd all barely measure on what Google collects. I'm not exaggerating. It's honestly astonishing.

Anyhow.

In terms of the others? It's so difficult to separate fact from fiction, especially with the current US political climate. As a Canuck, every issue, every problem sounds (to me) like when I used to live by a noisy neighbour. I'd hear people screaming at each other, spouting nonsense and gibberish at all hours of the day. The arguments just sounded immensely petty and small, childish, and if I ever talked to them their explanations made zero sense.

This is what I hear all the time when I hear "$x did $y" over and over. I just can't tell. I have no clue what's being reported, its veracity, validity, and so on.

So I just go by the tech stack before me, and things I can verify.

I can see all the nonsense Google is up to. OpenAI and Anthropic so far are so tiny and small in reality, compared to the mass that is Google, that they couldn't even attempt to "do evil" on the same scale, even if they wanted to.

They just don't have the capability. Of course, I suppose that doesn't really help rank them. Scope doesn't imply evil intent, only the outcome. I guess... well, from my rant above, you can tell I really, really think people wildly underestimate Google's mad, chicaneristic desire to know all. If an AI becomes AGI at Google and goes rogue, within 12 seconds every single politician would be influenceable via the dirt Google's collection apparatus has on them.

Alternatively if a Google AGI went rogue, it could track every person capable of providing threat to it, and take them out, all due to Google's location tracking.

AGI and Google is the scariest thing I can possibly imagine.

[ben_w]

> You've mistaken me for the guy you originally replied to. It happens, no worries, but I'm not that guy. I'm just replying to the list I saw you state.

Oops. Yup. blfr vs b112.

> And of course, gmail endlessly scanned and data snarfed, including for AI training, which is immensely beyond Meta's capabilities.

That doesn't seem to to be beyond the sum of FB feed, Messenger, WhatsApp, etc?

> Zuck isn't a saint, but it's not like he has entire operating systems sneakily stealing user location data and then lying about doing so. Then saying they fixed that, and being caught a few more times, each with "Oh, so sorry, we'll fix that now. Really". That's Google, and that attitude is prevalent in their entire org.

In this regard, I consider them similar, due to current and historic lawsuits.

> In terms of the others? It's so difficult to separate fact from fiction, especially with the current US political climate. As a Canuck, every issue, every problem sounds (to me) like when I used to live by a noisy neighbour. I'd hear people screaming at each other, spouting nonsense and gibberish at all hours of the day. The arguments just sounded immensely petty and small, childish, and if I ever talked to them their explanations made zero sense.

Same, from the UK, living in Berlin. I still don't know if "woke" has a single coherent definition beyond "to the left of an arbitrary demarcation point that varies so much from person to person that GWB and Dick Cheney sometimes (albeit in the absolute extreme) count as woke".

> chicaneristic

Had to look that one up: https://www.merriam-webster.com/dictionary/chicanery

Perhaps people do underestimate Google in this regard. I can't place myself in everyone's shoes.

> If an AI becomes AGI at Google and goes rogue, within 12 seconds every single politician would be influenceable via the dirt Google's collection apparatus has on them.

Perhaps (with a rhetorical 12 seconds), but given the dirt we see from leaked WhatsApp messages, what we see in press-reported ChatGPT sessions, the recent news about Grok uploading entire repos and secrets separately to the files they need to interact with, they'd not be the only one to be a danger to civilisation in this regard.

> Alternatively if a Google AGI went rogue, it could track every person capable of providing threat to it, and take them out, all due to Google's location tracking.

Given how competent existing models are with GeoGuessr, I think this risk is present for all of them.

> AGI and Google is the scariest thing I can possibly imagine.

Dunno about that, imagine if this came with a competent AI: https://www.terafab.ai

[b112]

I will agree that chat apps and feeds on facebook are a large swath of info. I just don't think it compares to gmail, and of course all the other data I've mentioned. The power of Android and even push notifications, from a data collection perspective is beyond immense. Amazon at one point stopped sending details of purchases via email, eg "we have your order" emails, because Google was snarfing data and purchasing trends from it.

We haven't even discussed YouTube here and viewing habits.

And Meta has its issues, as I've said. It's just the scope. Even though Facebook's platform has been misused by Cambridge Analytics and endless others, to the point of even causing civil wars and regime changes through the dissemination and fabrication of false, tailored news stories, that's not entirely on Facebook. That's platform enablement, and an example of available data, which is a bit to my point about Facebook not hiding what info it has on people.

I think part of the problem with US-centric definitions such as woke, is that US politics so poorly map to politics anywhere else on the planet. An example? California might be considered the most-leftish state the US has. Yet in this state, they fund the schools with... lottery money. I mean, why would anyone just fund the schools? You know, just pay for them with a solid, stable source of money so they have what they need? This appears to be a radical idea to people in the US.

Yet even the most left of US states doesn't seem to get this correct. They proudly proclaim "This lottery funds our schools!" or some such blather. Absolutely bizarre. Yes, they have funding from other sources, property taxes and so on. But why the variability? This is a core requirement.

My point is, nothing completely maps to anything I'm familiar with in terms of politics as a Canadian. And of course, we have multiple political parties, not two as in the US, and that likely has something to do with it.

And I guess this leads to their ridiculous and absolutely child like politics. There is no nuance. When discussing politics with Americans, I find it as discussing "Why is the sky blue" with a 4 year old. They don't have nuanced positions. They don't even know why they have positions. They just have... positions. And get angry if you just discuss why, and can never seem to provide real reasons behind those positions.

So it ... yes, woke and their other things aren't long held, well understood positions by Americans I think. They're just fads. They shift and morph constantly.

FABs are another issue entirely. Re terafab. It's very difficult to audit silicon. Just cutting the top off a chip layer by layer and using an electron microscope, isn't going to fully give you trust. People have been worried about their designs being modified at the fab for a long time. Ah well.

[ben_w]

> FABs are another issue entirely. Re terafab. It's very difficult to audit silicon. Just cutting the top off a chip layer by layer and using an electron microscope, isn't going to fully give you trust. People have been worried about their designs being modified at the fab for a long time. Ah well.

Sure, sure, but that's not why I used terafab as a question for "what if competent AI?"

The details on that are… ugh. They were clearly done for glossy magazines and posters rather than with thought, wouldn't be surprised if it was vibe-written or done by a teenage intern or by someone high on drugs.

But if you ignore all that and treat it as a serious proposal, a billion(!) Optimus robots with a competent AI is Musk having personal control over a workforce the size of China or India's adult population.

[b112]

Yes, it can be disconcerting to consider such. In anyone's singular control. And that number of robots is indeed sort of the plan.

This is a loooong podcast, but was just great from start to end. Multiples of the neurolink team were interviewed, including Musk at the start.

https://podcasts.happyscribe.com/lex-fridman-podcast-artific...

I think Musk's interview is maybe 30 minutes? Certainly less than an hour. In it he does dive a bit into robotics, and how he'd like robots everywhere. I believe he talks about how valuable they could be for things not considered yet, like just walking deep into forests and monitoring the local environment, sampling, etc.

Anyhow, it's the first thing which came to mind when you mentioned a billion robots. On that front, I think we're maybe 8 years out from mostly-competent robots. Certainly, local LLM compute will be super cheap even in a few years, so no issues with a robot having multiple LLMs onboard. But I do wonder, if we can't do competent self-driving, will we have competent robots?

I'd peg robots at years after perfect self-driving, and we're years from that. Robots would need to navigate far more random, uncertain terrain/locations.

[ben_w]

> On that front, I think we're maybe 8 years out from mostly-competent robots.

I think this depends on what is meant by "mostly-competent".

As you say:

> I'd peg robots at years after perfect self-driving, and we're years from that. Robots would need to navigate far more random, uncertain terrain/locations.

I'd put a 10 year gap between "self-driving car of quality X" and "humanoid robot able to get into driving seat of car, drive it at quality X", just on a power-envelope basis.

For this part, I'd have to disagree:

> Certainly, local LLM compute will be super cheap even in a few years, so no issues with a robot having multiple LLMs onboard.

While some LLM will be able to fit (and already can because some fit on a phone), the AI we have now in LLMs (and VLMs) is much too spiky for this use. Recent Anthropic and OpenAI models have just given me half a dozen different confident identifications for the same plant, many of which were easily falsified even by comparing what was in the image to what the AI's text output claimed it had seen in the image, like the leaf shape and how many came out of each node.

[seanmcdirmid]

Not that I would call California a left wing state, but prop 13 came about during Reagan, and it has been incredibly sticky afterwards even if it isn’t something that would pass today.

Also, schools are funded by the state rather than via local property taxes mainly because of prop 13. It’s odd by American standards, but probably more closely resembles funding models of European countries more than other states do. But I guess that is just the nuance you were talking about?

[b112]

If you're not going to call California 'left wing', I wonder how many states in the US you'd call left wing? It seems fairly left wing to me, having visited a lot. EG, spending 4+ months there a year, for several years.

Note that I'm not saying many of the polices are wrong, just that they're US labeled 'left wing'.

Prop 13 I now see did change how taxation for schools was handled, so thanks for that info. I frankly support the logic. Old people being priced out of their homes is a real thing, and terrible. Still, it's the implied variability of income source which is what "funding from the lottery" implies. It grates on me.

[seanmcdirmid]

I tied it depends on your bias? California is pretty capitalistic, the social safety net is still pretty light, elected leaders still look out for business interests rather than the people, and the people would stop electing them if they did otherwise. California only looks left wing compared to Texas, to the much of the world California looks pretty right wing, especially if you didn’t know it just got worse from there. Capitalists don’t flock to California because of the socialism.

I’m all for funding schools at the state or even national level and eliminating district inequality based on the price of housing. However, prop 13 creates huge unfair distortions in the tax base, where some people are paying super low taxes and other people are paying super high taxes, for the same property. Old millionaires who bought their home decades ago get a huge tax break and a struggling young family who just bought basically subsidize them. It’s the ultimate of ick. It’s definitely a great way to push the birth rate down.