@BlackHC You’re right, in fact, we should hope to see monopolies everywhere! Winner-takes-all scenarious have historically proven to be the best outcome for everyone! When VW dominated the EU car market they were also motivated and able to preserve ecological standards better than ever.

Being very transparent and honest with my conclusions after thinking about this 😊

Google DeepMind is a great place, especially the science and foundational unit historically, but it is not an AGI company because it is not an independent company

It fully belongs to Google, and Google very much is not an AGI company as far as mission and governance are concerned

The sooner one accepts that the better 😇

@BlackHC Giving a company monopoly and believing they will do good is probably not going to work out as you imagined either...

@BlackHC Yes. https://open.substack.com/pub/hollyelmore/p/anthropic-delenda-est?r=15m0xc&utm_medium=ios

@mtavitschlegel Not sure this is comparable. What are your thoughts on the coordination issue in a race scenario? Do you have any?

The point was that if you have a strongly competitive race, it is more likely that all skimp on safety vs if one company has a lead and also strongly invests in AI safety etc

RE board: you mean like Google? Yeah that is a risk but then Anthropic does have better governance in place and it is for benefit and not for profit. So overall it is more trustworthy compared to OpenAI and Google right now

@BlackHC I thought you worked for Google, are you looking to switch to Anthropic? Or am I missing some layered sarcasm in your message?

@BlackHC @Zergylord The reason Google is not an AGI company is that it doesn’t need an IPO and rounds of fundraising. We don’t need AGI for AI to enable scientific breakthroughs at the current pace.

@Zergylord I think quiet quitting is cruising at L5 without a Twitter account and working a 36h week while also investing in OAI and Anthropic (once IPO'ed) 😅

@BlackHC Anthropic is literally the most hawkish of all the labs save xai. They were the first to collaborate with palantir and therefore the DoD. Their models have provably been used to kill people.

@BlackHC Is this that quiet quitting thing?

@RBH_Research What does "giving" mean here?

@BlackHC What's 'helping Anthropic win the race' while at Google called them?

@BlackHC Imo it’s dangerous if a single entity has absolute power over whatever and is in a position to decide what’s best for everyone. This is why we have democratic institutions & free markets, which both historically work best when we avoid (power/economic) centralization

@UltraTopSecret I'm not sure I'm really really great, so it doesn't apply to me 🥲

@BlackHC I kinda disagree with the premise that we need that. We also don't rely on coordinated self-regulation eg in aviation or pharma? For this we have democratically controlled third parties. But if one comp dominates, you get information asymmetry, no peer comparisons etc. Also, what

I'm not arguing against democratic institutions and free markets

I'm worried about 3-5 companies with a mix of bad governance and a lack of alignment and safety focus cracking RSI while democratic institutions are not able to keep up

I think it is a much better scenario is one company leading comfortably with a wider base of companies trailing incl open source models which serve to squeeze the margins and slow down frontier research funding

That is much better than 3 companies at the frontier (and me not trusting 2/3 very much because they're playing catch-up)

@BlackHC @mtavitschlegel “it is more likely that all skimp on safety vs if one company has a lead and also strongly invests in AI safety etc”

This shows a level of credulity to your own propaganda and ignorance of history that I find completely baffling. Do you genuinely believe that this is true?

@BlancheMinerva @mtavitschlegel You disagree that they're strong on AI safety and alignment research?

Good point! Responsible business practices usually require regulations for profit maximizing companies

I guess the question is whether having binding but otherwise untested legal obligations as part of the setup is sufficient. Ie are Anthropic and OpenAI's setup sufficient to self-regulate?

And these governance forms are relatively fixed and not adaptive, so will they fail over time? It requires continued goodwill to adapt them