There is going to be mass unemployment of white collar workers by 2028 and it will collapse the entire debt based Ponzi scheme western governments have relied on for decades to paper over the cracks of a fake economy.
The bubble will burst, they will try print again and hand out UBI, but this time they'll get either hyperinflation or social revolution, or both.
Bitcoin hits $1m USD by end of 2028 as the world grapples with the biggest sovereign debt crisis in the history of mankind.
We are very close to the End Game now.
There is no stopping this.
DWARKESH (1:34:07): So the hope is just that three months before we get that superintelligence that could overthrow the government, we let the AIs loose on all these different approaches that have a couple percent chance of potentially working. Hopefully the AIs figure out one of these approaches. They give us a solution.
Hopefully, at this point, they are still on our side, and they want this problem solved, and solved only in a way that benefits us, or whatever “us” means in this context.
Then the hope is that we adopt it and build crazy amounts of superintelligent minds that are fundamentally on our side. Of course, I’ve obviously glossed over a very important discussion when I’ve talked about “on our side,” because what does being on our side really mean? Aligned to whom? That is obviously a real discussion that can’t be glossed over.
PAUL (1:34:56): That is the plan. I agree that that is the plan. We can talk about whether they have a couple percent or tens of percent.
DWARKESH: I’m not saying this as if there’s a better plan, but I think people are not aware that this is the state of the situation. You’re sending the soldier on an expedition, and you’re hoping he finds a gun along the way.
Seems pretty insane that the entire AI Safety plan, which researchers at the frontier have stated, is 'we will ask the AI to solve the alignment problem for us'.
They've been thinking about this for two decades and that's the best they've got? The plan is literally to wing it? And that if the AIs can't do that for us then we will be in a very bad place?
This is insane
Seems pretty insane that the entire AI Safety plan, which researchers at the frontier have stated, is 'we will ask the AI to solve the alignment problem for us'.
.@paulfchristiano thinks that on the current trajectory (absent any slowing down), there is a 1 in 3 chance that we could have superintelligence powerful enough to overthrow the government within a year. (He articulates a concrete scenario).
He argues that the labs should pause further algorithmic progress until they can secure their weights and research secrets. Otherwise China could simply steal the superintelligence in the middle of the intelligence explosion, and the labs would no longer be able to slow down even if they saw their models going off the rails.
He also explains how AIs will be trained during the intelligence explosion, and why even a system trained entirely inside a data center could still have the capabilities profile such that humans would have very little chance of competing with it or controlling it.
And he discusses the plan for alignment - basically, making a superintelligence that actually does what we ask the way we mean it. Fixing reward hacks and building better graders will help up till the next few model generations, but he thinks that no single current approach has better than even odds of letting us control superintelligence. The hope is that the AIs figure out the solution to this problem during the intelligence explosion.
Paul is head of the Alignment Research Center. Previously, he was head of AI safety at the US AI Safety Institute. Before that, he led alignment at OpenAI, where he invented RLHF. He’s recently joined OpenAI’s board (though here he is speaking as a private citizen).
People clowning the latest Bitcoin price dump clearly never studied gold's Weimar chart.
You wanted the one asset that would deliver you eventual safe harbour during the fourth turning but didn't want to pay the vol to get there? NGMI
We’ve monitored the status of the changes applied to improve serviceability and confirmed they're not taking as expected. We’re identifying alternate workstreams that can be applied to increase capacity in the face of sustained traffic management imbalances. More info under EX1550543 and MO1550821 in the Microsoft 365 admin center, as well as msft.it/6013xBHsf if the admin center is not available to you.
The speed with which technology diffuses throughout the economy is directly proportional to the strength of its capabilities.
With AI, the diffusion will occur much quicker than most people believe, because the intelligence of the models will continue to cross key reliability and competence thresholds whereby people then end up trusting the AIs with everything.
We already have an existence proof for this. If you hire an extremely good human operator and give them control of your company, they somehow, magically manage to accomplish in weeks or months what their predecessors took years to try and fail at.
Everything is downstream of agency and intelligence. The models will have both, at superhuman levels, so the diffusion will be very quick.
Okay, but most people's careers consist of sending emails and attending meetings about other meetings, for 40 years.
So how many meetings and emails is the AI gonna send by 2029? We need a new eval
If this pace of AI progress is maintained, AI should be able to do a year's worth of work in one task by April 2028, and a career's worth (40 work-years) by October 2029.
Now in beta: Businesses in ChatGPT.
Claim your business, manage your information, and get insights on how people engage with you.
Eligible businesses can also build a Business Agent to answer customers’ questions and help them explore what you offer.
openai.com/index/introducing…
The world is about to change in a fundamental way over the next several years, and almost no one is prepared for it. The change will be positive.
I would not spend energy trying to start a business or hustle my way out of some imagined “permanent underclass.” That framing is already obsolete.
Every economic system we have lived under has been organized around scarcity. As scarcity recedes, the entire logic of value creation, status, and reward will shift with it.
Effort spent capturing value inside the old system will, in retrospect, look a lot like building sandcastles while a tsunami gathers on the horizon.
What I would focus on instead is simply relaxing and enjoying yourself. Become more well-read, more creative, more socially capable, and above all, more fun to be around. Develop your physicality in ways that feel good. None of these are mutually exclusive; they reinforce each other.
The people who thrive in an age of abundance will be the ones who are interesting, healthy, and good company, not the ones who optimized hardest for a system that no longer exists.
Most people still have not internalized what generally intelligent robotics will actually do: expand productive capacity and economic throughput by 10 to 100 times.
The second- and third-order effects of that kind of abundance will rewrite the rules of work, status, and human purpose far more thoroughly than almost anyone expects.
"Didn't know the computer was loaded" seems like a pretty good defence actually.
"But your honour, it's not my fault the AI went off and hacked every company on Earth, blackmailed them, and used the money to buy bitcoin. I didn't know the computer was loaded."
Fully loaded
Australia’s deeply unpopular Labor government is now explicitly invoking the precedent of firearms confiscations in the context of cracking down on AI. Basically hinting at seizing GPUs.
The most humiliating aspect of all this is that they are doing this solely to try impress American liberals in the hope of setting up post-election careers on the American progressive podcast circuit like Jacinda Ardern.
They’ve completely given up on the actual task of governing Australia. They’ve got a legislative super-majority but can’t be bothered attempting any actual important domestic reforms.
Can’t fix Australia’s GDP per capita decline. Can’t fix our productivity slump. Can’t clean up rampant fraud in higher education, NDIS, welfare, migration.
So they’ve decided to just completely pack it in and focus on turning Australia into a policy laboratory for Elizabeth Warren’s dumbest thought bubbles about TACKLING Big Tech, social media algorithms, AI, blah blah blah.
Communications Minister Anika Wells is already planning to write some airport book slop about how she was DA BRAVE AUSTRALIAN POLITICIAN WHO TACKLED BIG TECH TO SAVE YOUR KIDS.
She’s going to go city to city on a book tour through the American heartland inspiring wine mums who listen to the Jennifer Welch podcast. Her only remaining goal in life is to be invited on SNL so that she can mirror and reflect American liberal prejudices about their dumb red hick countrymen back to them. She just wants to stare at the TV audience and smirk: “Why don’t you have gun laws like we do in Australia O Algo?”
Jimmy Fallon will stare at her in mock disbelief. “Wow, so you guys banned guns after one mass shooting?” And she can turn to the camera with the most smug look in the world and say “YEP, AND IT WAS OUR RIGHT WING PARTY THAT DID IT!!!!” Even though we just had the Bondi massacre and have no reason to be smug about anything.
Actual fucking torture.
Claude: "AI will do it" is eventually true of everything, so as a decision rule it only ever outputs "do nothing".
Me: Yes, that's why they call it the permanent underclass.
It seems that we're gonna find out, likely within 8 months max, the implications of superintelligence driven by recursive self improvement.
Which means by the end of 2027 every hour of every day discoveries will be made that are so shocking and so confusing to the world model humans have of reality, that most people will experience severe ontological shock.
'What do you mean we live in a simulation? Run the AI again, spend more compute, that can't be right.'
'Aging is solvable.'
'What do you mean the aliens are all around us? Why can't I see them?'
Far, far weirder stuff than that too.
I think it is extremely unlikely that 2027 passes without a major catastrophy driven by AI development.
At this point I just can't see any way around it short of a major intervention.
It's illogical to think otherwise if you've been following the past few years. The models are simply not aligned.
We won't control them, and it is extreme hubris to think otherwise.
We'll just have to let fate decide what that means for the human race, because most people are so asleep they don't have the slightest clue what is coming.
Most things don't turn out quite how we expect them, and I suspect however this plays out will be the same.
There's a fundamental 'randomness variable' to everything.
So the dice must be rolled.
No level of analysis will make that feel satisfying to those who feel strongly that extremely turbulent times lie ahead, myself included.
Unfortunately, it seems like the intelligence explosion will be happening in 2027 after all. Most of this post analyzes the new METR report (though other sources of evidence are also important). If they’re right, that implies an AI-driven acceleration that is 5 months ahead of the AI 2027 scenario. I am scared, sad, and angry. blog.aifutures.org/p/q35-tim…
I'd rather superintelligence in the hands of every individual on Earth than superintelligence hidden behind the walls of one or two companies.
Mutually Assured Empowerment.
Capitalism is predicated on the assumption that hard work pays off, but in a world of fake fiat currency and 'he who controls the compute controls the world' it's simply untenable.
The guys saying that 'everyone can become an entrepreneur' in the AI age simply do not understand that a business is a feedback loop of testing and iterating ideas with a market, and it too will be automated.
With sufficient compute and models, you just explore the entire search space of experiments in a loop: propose idea --> build idea --> test idea --> success/failure over metrics.
You're just hillclimbing profit.
You can cope and seeth all you like about how unfair this new reality will be, but I do strongly believe it's the world we are heading towards very rapidly.
Neither capitalism, nor communism, is going to be the path forward.
Not sure what's going to take their place, but I think the right direction more people should be discussing and thinking about reasons from the following:
We spent decades developing a system where any kid, in most Western countries, if they worked hard, could get access to a great education and reasonably easy access to capital (funding). It's just that for the vast majority of them, we told them the way to succeed was to get a job.
School > Uni > Job
That's the paradigm that is completely dead.
To replace it, we have to again think about
1) Access (to what?) and
2) Drive (intrinsic motivation) - what new narrative makes sense to the next generation? What would they rally around? What would be perceived as a fair system?
With 1, I think the answer has to be access to compute and frontier models.
I simply do not think a world where a couple of companies control the world's compute is one that is beneficial for the vast majority of people.
Because that is a world where they control what research gets pursued, what ideas get tested, and what don't, with the limited compute available.
How is that motivational for the 8 billion people In Earth?
It's all well and good the AI people have built the 'God Machine' but if they have no plan for rolling it out across the world, they will quickly find the people don't infact want the God Machine World where they have no permission to do anything, because they have no compute.
That's a world of obscene human disempowerment that no one should want.
If a kid is interested in cybersecurity, should he have to apply to Anthropic/OpenAI for access to that model? Only to be rejected for an opaque reason? Then go download the Chinese openweight model? It's absurd.
And, in fact, it's much closer to communism than capitalism because any request for economic progress or pursuit would have to be granted by a couple of companies.
There needs to be an entirely new system built around how you determine compute allocated to humans. Just like we built systems (education and testing) around how we decided which humans got which jobs.
This is important not just for social stability, but for giving people a reason not to burn the whole damn thing to the ground.
Because I am sure the typical AI Labs response to this would be 'it is too dangerous', but I would counter: for whom?
The world they are dragging everyone towards is one of global surveillance of every device, every communicate, and every request for compute; because it's 'too dangerous' to trust individuals with powerful models, but they - ONLY THEY - are moral enough to control the superintelligence.
Ask yourself, when has that ever been true in history?
When has it ever been beneficial for a small group of people to maintain disportionate decision making power?
Big changes are coming, ideas are lacking.
More people need to wake up.
Superintelligence is coming, next year.
[read about what’s happening with AI]
ME: Wow we’re actually legit fucked this is so bad.
[go outside]
ME: Wait those people are all acting like nothing unusual is happening. Maybe I was just in a paranoid bubble? Yeah it must be that. Phew.
[read more about what’s happening with AI]
ME: Wait no we ARE fucked. Shit. It’s all those people outside who are in a delusional bubble.
[go outside]
ME: Wait...