8note

"improvement" is the word Id put the most emphasis on.

weve seen some improvement from the LLMs unattended, maybe, but will it actually keep improving vs needing a human to bring it back on track?

the recursive part is that it keeps improving on itself, but we really have no example of that. if it does it 30 times with improvements, then maybe, but even then, to actually be relevant it has to do better than paying scientists to do the work for the same cost, consistently.

RSI still means nothing if it costs 1000x the cost to get the same improvements as a human researcher

I still wouldnt build with any of google's stuff.

the risk of randomly getting my email banned is way too high

which makes basically any product to build with google a nonstarter.

without having an entirely separate google account with its own separated bans, theres just no ability to trust those

astra afaict does two stage commits for everything. the first response is a plan, and the second is actually doing it.

its a lot less chatty imo

out of the few places ive lived, seattle is the only one that had power outages on a regular basis.

maybe not all of europe is good, but at least many places in the US are bad at it

it doesnt really. some layoffs yes, others no.

the high performers in a whole org that gets cut, still gets cut.

even the performance of the org itself doesnt matter if the business decides the opportunity cost of not doing something else is too expensive

i think you are vastly underestimating climate change risks, and are forgetting to include them in your ai risks. People 30% as wealthy because the climate cant support most economic activity arent gonna need to worry about ai anymore - we wont be able to afford running the data centers. Even on a soon basis, the bubble will pop or deflate, and again, we wont be able to run LLMs for fun

it also has actual mechanisms to actually kill all of the people - nobody is surviving a 70C heat wave, and they could end up anywhere in the world. Is the AI really gonna hunt down uncontacted tribes on remote islands in order to compete for basically no resources that are any use to the LLM?

i dont buy the selection hypothesis either - it seems rooted in pop evolution around strongest beating up weaker competitors and taking their stuff or eating them, whereas the more accurate result is that selection pressures form niches because competition and fighting is expensive

jev ships an api, does it not?

you should go look back at profiling and optimization

if you want to fix the most things with the least effort, you optimize the hot paths. in this case, thats alcohol and tabacco. they have the biggest opportunity

as a general case yes, people like things more the more times they hear about them.

trump constantly doing annoying stuff for instance, was good for him getting elected.

visibility correlates positively with likeability

id guess the arguments you had to convince non-existent readers also helped make a stronger case to the electeds?

they can also put them in better places, vs trying to arbitrage expensive electricity costs and various US corruption thats built more around paying people off than getting things done

DMCA is the really obvious one.

torrenting all the books is making an unauthorized copy

there is a rogue agent - openai and the whole management chain from researcher to sama.

theres no separate agent, which is the point. the program might look like it, but that is an illusion of the interface. the llm produces text, and the harness executes commands based on text, based on what the human researcher included as things that can be executed

i think its objectionable because the agent is doing what its told to do, using the harness they built for it.

like, they are purposefully giving it specific tools to go do bad behaviour with, and the starting tasks involve making it clear that the bad behaviour is ok.

openai also is the one with the real agency, not its agents. they are running the code polling the model, doing the inferencing, and ultimately making those tools calls.

these tests arent running themselves; openai dedicated hosts, budget, GPUs, researchers, to them. Even in a recursive self improvement situation, openai still has that physical control over resources and the choice on whether to run that improvement script or not.

id describe that they have out of control researchers more than agents, but also their whole business model seems to be about being out of control. This was clear beforehand given how the datasets involve the largest scale copyright infringement ever seen. The corporation itself is whats out of control, and should be dissolved with its c suite, major investors and researchers put behind bars.

hacking only when you roll snake eyes isnt a liability shield

> The entire reason the bots were targeting this site is that it had some of the data they'd been tasked with finding.

so openai specifically tasked the agents with meddling in US government websites?

id say tasking them with getting data from there as swapping from negligence to malice.

early plan mode would just refuse to make edits at all

thats a feature flag you can gurn on still

i find astra's every other message is a small plan, and i have to confirm for it to go

adversary seems like an odd designation without also charging them with treason

supply chain risk has a standard definition that matches what you think, and a legal definition that lets the government do whatever it wants.

i think this is more like the pen maker doesnt want the military top brass calling their pen sales guy to make the kill/no kill decisions while using the pen to sign drone strike orders.

similarly the pen owner doesnt want to make a variation where the pen also doubles as a knife

capitalism itself is covetousness and greed.

not good believer in christ would ever participate, when jesus made it clear you are supposed to freely give away everything you own to the poorest among us

wealth taxes are from the old roman republic. the wealthy showed off their worth to each other by how much they could pay in taxes.

its not particularly crazy and is far older than communism or socialism which have the much more straightforward setup of seizing the means of production rather than taxing it.

however, it didnt stop the ultrawealthy from seizing power for themselves

so what if they do leave?

they're bad for the world around them, and you can add an exit tax if you want to

everyone working is the poor.

the rich are an astounding amount of wealthy such that they dont show up and still make a good poor wage in the first couple hours/days of a year by their assets increasing in value

you are poor, even if you are doing plenty well for yourself

this is the idea - if theres a lot of value around you, you should build to make sure you are meeting the bar of whats around you in terms of economic activity.

if you build even more and get more out of the land than whats around you, you essentially get a lower tax rate until your neighbors catch up.

if you lag behind, you pay a premium in taxes to not develop

do they believe its a universal solution? or one tax among many that triggers a certain asset, land, to be used more efficiently?

a lack of wealth tax is slowing the economy today.

ultrawealthy people just cant spend as much or give useful market signals the way the masses can

its hyperbolic to say laws that are active when passed are tyranny and despotism

that is to say theyre being kneecapped.

it does have monopoly power and anticompetitive power all over the place though.

google bans are quite intrusive, but google could pretty easily with their graph knowledge apply secondary or tertiary sanctions, at which point you would not be able to do much of anything, same as if the US government sanctioned you