2025-01-29 19:30:37
AI is progressing so fast that its researchers are freaking out. It is now routinely more intelligent than humans, and its speed of development is accelerating. New developments from the last few weeks have accelerated it even more. Now, it looks like AIs can be more intelligent than humans in 1-5 years, and intelligent like gods soon after. We’re at the precipice, and we’re about to jump off the cliff of AI superintelligence1, whether we want to or not.
When are we jumping?
What’s at the bottom?
Do we have a parachute?
Six months ago, I wrote What Would You Do If You Had 8 Years Left to Live?, where I explained that the market predicted Artificial General Intelligence (AGI, an AI capable of doing what most humans can do) eight years later, by 2032. Since then, that date has been pulled forward. It’s now predicted to arrive in just six years, so six months were enough to pull the date forward by one year. Odds now looks like this:
But AI that matches human intelligence is not the most concerning milestone. The most concerning one is ASI, Artificial SuperIntelligence: an intelligence so far beyond human capabilities that we can’t even comprehend it.
According to forecasters, that will come in five years:
Two years to weak AGI, so by the end of 2026
Three years later, Superintelligence, so by the end of 2029
As you can see, these forecasts aren’t perfectly consistent, but they are aligned: An AI that can do any task better than any human is half a decade away.
Who’s best positioned to know for sure? I assume it’s the heads of the biggest AI labs in the world. What do they say?
Here’s another quote from him:
I think powerful AI2 could come as early as 2026.—Machines of Loving Grace, Dario Amodei, CEO of Anthropic
And another:
Two years ago, we were at high school level. Last year we were at undergrad level. This year we’re at PhD level. If you eyeball the rate at which these capabilities are increasing, it does make you think that we’ll get there in 2026 or 2027.—Dario Amodei with Lex Friedman
Moving to OpenAI, the maker of ChatGPT, here’s its CEO, Sam Altman:
We are now confident we know how to build AGI as we have traditionally understood it. We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word. With superintelligence, we can do anything else.—Reflections, Sam Altman
Here’s the head of product:
"AI going from the millionth best coder, to the thousandth best coder, to the 175th best coder in 3 to 4 months… We’re on a very steep trajectory here, I don’t even know if it will be 2027, I think it could be earlier.”
Why does he claim that AI is now the top 175th best coder in the world?
Note that coders are the ones actually developing AI. So if AI is now the 175th best coder, then AI can code AI better than most humans themselves.
Here’s a researcher at OpenAI:
And another:
We believe o1 [the model OpenAI released a few months ago] represents a new scaling paradigm and we're still early in scaling along that dimension.—Noam Brown
Everything that we’re doing right now is extremely scalable. Everything is emergent. Nothing is hard-coded. Nothing has been done to tell the model: “Hey, you should maybe verify your solution, you should back-track… No tactic has been used. Everything is emergent. Everything is learned through reinforcement learning. This is insane! Insanity!
Since joining in January 2024, I’ve shifted from “this is unproductive hype” to “agi is basically here”. IMHO, What comes next is relatively little new science, but instead years of grindy engineering to try all the newly obvious ideas in the new paradigm. Maybe there’s another wall after this one, but for now there’s 10xes as far as the eye can see.
This is what Biden’s White House had to say about this:
It might be why Sam Altman has scheduled a closed-door briefing for U.S. government officials for Jan 30. From Axios:
AI insiders believe a big breakthrough on PhD level SuperAgents is coming." ... "OpenAI staff have been telling friends they are both jazzed and spooked by recent progress."
If you think this is only hype from AI leaders seeking new investors, consider the words of Demis Hassabis, recent Nobel Prize winner thanks to his AI work, and CEO of DeepMind (part of Google, so no need for fundraising):
There's a lot of hype in the area. [AGI is] probably three to five years away.
But is there tangible proof that these AI models are getting so good?
AI has a history of taking time to pass humans at different tasks.
But tasks that used to take decades for AI to surpass human performance now take months. This is how fast things are moving.
OpenAI’s cutting-edge model, o3, has not been released yet, so let’s look at what people using the much dumber o1 had to say about it.3
According to a MENSA test, it has an IQ of 120.
To give you a sense of how intelligent that is, it would place o1 between superior intelligence and very superior intelligence, and puts it at 87th-91st percentile intelligence. In other words, this thing is more intelligent than 9 humans out of 10.
Think
about
that.
Here’s one way to provide some intuition about it (try to figure out the answer to the question below by yourself!):
So what is your answer?
Think about it!
OK here it is:
I corroborated this independently.4
It certainly took me more than 5 seconds to figure this one out, and I’m going to guess many people don’t solve it at all.
In July of 2024, Google DeepMind’s AlphaProof and AlphaGeometry nearly got gold in the International Mathematical Olympiad (IMO):
o1 is a generalist. It’s not specialized in math. And yet it scored 83% in a qualifying exam for the same IMO. The previous ChatGPT scored 13%...
Why does this matter?
Reaching gold in the IMO is a threshold for AI researchers, which they interpret as “being able to reason”.
Paul Christiano—inventor of RLHF, certified genius, and arguably the most respected alignment researcher in the world—previously said he'd update from 30% to 50% chance of a hard takeoff if we saw an IMO gold by 2025.—@AISafetyMemes
As a reminder, a hard takeoff is when AI gets intelligent enough to improve itself, and so it does, becomes more intelligent, improves itself more, gets more intelligent, and FOOM, becomes superintelligent in a matter of months, weeks, or months through its recursive loop of self-improvement.
If you want a sense of that, listen to the CEO of NVIDIA—the company making over 100%5 of the profits in AI today:
In 2021 Paul thought there was only an 8% chance of this happening! He was famously the most well-known slow-takeoff advocate. Fast takeoffs are scary because they're way less predictable. Human institutions have almost no time to adapt to insane rates of change.—Source
Look at how it performs against PhD-level science questions:
It responds better than PhDs!
Here’s the broad battery of tests:
Also, recently some AIs have received a 100M token window (meaning they can consume about 750 novels to give you an answer).
All the stuff I just told you is about o1, the previous model! What do we know about the current model—o3, the one OpenAI is testing and whose researchers are excited about?
The GPQA is a difficult science test designed for PhDs and impossible to solve by Googling. PhDs get scores of 34% on average. In their specialty area, they get about 81%. The new OpenAI model, o3, gets 88% right.
In other words, o3 can solve problems better than PhD experts in their field.6
Another benchmark is FrontierMath, a set of rigorous, unpublished math problems designed by mathematicians to be incredibly difficult to solve.
These are genuinely hard problems… most of them look well above my paygrade.—Evan Chen, International Mathematical Olympiad Coach
AIs never scored higher than 2%, but o3 got 25% right.
ARC-AGI is a famous test of fluid intelligence, designed to be relatively easy for humans but hard for AIs. While the average human gets 75% and graduate STEM students get 100%, o3 beat all previous AIs and got 87.5%.
An AI expert predicted two years ago (in 2023) that it would take 2-3 more years from now (so 2027-2028 or so) for AI to pass 70% at ARC-AGI…
Let me summarize this:
OpenAI’s o1 model was so good it was already quite elite.
Their new model, o3, is so good that it’s better than experts in their own fields.
The progress between these two was a matter of months.
AI researchers say this isn’t even optimized!
They are so flabbergasted at this progress, that many conclude we will reach AGI in 1-3 years.
Are you in awe yet?
But maybe they are missing some barriers that could hinder this growth? Can intelligence progress towards AGI hit a wall?
Maybe the technology has some fundamental limit to the types of problems it can solve? No:
We have mathematically proven that transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed.—Denny Zhou, Lead of the Reasoning Team at Google DeepMind
So it looks like technology and intelligence won’t be the limiting factors to ASI. If anything, it will be the raw materials of AI. These are:
Data, to train the AI models
Compute capacity (that is, chips), to train and run the models
Electricity to power the computing
Money to afford the rest.
Algorithm quality, to optimize all of the above
It looks like (4) money won’t be a problem.
The eyes of the world are on AI. Everybody knows it’s the future, and are willing to put in the money to make it happen.
Companies like OpenAI, Amazon, Meta, and Google are seeing how much electricity they will need, and are investing to build their own capacity—so much so that they’re all building or reopening nuclear reactors, which provide stable electricity. Their higher price vs fossil fuels or renewables doesn’t matter to them, because electricity is a much smaller share of their costs than compute.
How hard will it be to produce all the electricity we need? Let’s take the US as the most relevant example:
Data centers already use 4.5% of electricity demand in the US, and that will probably rise to 7-12% in four years. How hard will that be? Well, the US hasn’t grown its electricity generation by much in the last couple of decades, so this might be a stretch.
But renewables are exploding:
US data centers need ~200 TWh of additional electricity by 2028, but solar and wind have been adding ~200 TWh of electricity per year, so I’m not really concerned about electricity.7
Compute is probably going to be the limiting factor. That’s why NVIDIA is the 3rd most valuable company in the world, and why two other chip companies (Broadcom and TSMC) are 10th and 11th.
But there is so much demand that new suppliers are competing with NVIDIA more and more, like Cerebras, AMD, Intel, Google’s TPUs (with so much internal demand that Google consumes all its own capacity and doesn’t sell these anymore), Microsoft, Groq…
It’s true that we’re hitting a wall here: We’ve already fed these models nearly all the Internet. The amount of content that isn’t digitized but can easily be retrieved might double the available data, but that’s it. So? There are two answers to this.
First, synthetic data. We already have physics engines that can create realistic environments that AIs can navigate. For example, they can help train self-driving cars. A more exciting source of synthetic data is simply to use AI to create data to train other AIs. Why does that work? Isn’t that garbage in – garbage out? No, because in many scientific fields, it’s easy to create exercises, get an answer from AIs, and test their accuracy. If the result of the AI-generated exercise is right, the AI is told so. If it is wrong, it is also told that and learns for next time.
Second, human brains don’t need to consume the entire Internet to be smart. This means we’re vastly overshooting right now: We’re using trillions of tokens of data because our AI training and computing algorithms are not very optimized yet. Nature stumbled upon the algorithm for our brains through trial and error. We can at least approach this efficiently.
In fact, we just did. A little known Chinese company has released a ground-breaking AI that is causing everybody to rethink the world of AI.
DeepSeek is a small Chinese organization that put out a new AI model called R1, which you can try here on desktop.8 This is its “AI quality score”:
It costs 25x less than o1:9
And who is behind this? A 100-person Chinese startup with zero VC funding.
This is an open source model, too, meaning that everybody can look at the code, contribute to it, and reuse it! They gave it away. All the tools they employed to make this breakthrough are now available to everyone in the world.
And DeepSeek is already improving itself!
How is Meta (Facebook) reacting to this? Meta began releasing versions of its own open source model, Llama, two years ago. Llama is worse than R1, and way more expensive!
And DeepSeek is not the only one doing this. ByteDance Doubao-1.5-pro matches or exceeds OpenAI’s GPT 4o and Anthropic’s Claude Sonnet 3.5 benchmarks and is 50x cheaper.
Now let me tell you: I just tested DeepSeek for the first time, and my experience has been profound. For the first time, I feel like I have a capable collaborator for the type of research I’m doing. I am dumbfounded.
For years, I have been wondering why a certain country is so populous. I had asked AIs many times, but they always failed me:
They didn’t consider root causes, like “There’s a lot of people because there was a fertility boom.”
They mixed causality, for example, saying things like, “There’s a lot of people in this country because most people live in cities, and cities are very dense”.
When I made mistakes, they didn’t notice, so they couldn’t question my assumptions.
They proposed a bunch of bullet points, without considering the relationships between them.
And they made many more mistakes.
With DeepSeek, for the first time, an AI was able to independently come up with the same hypotheses as me, without me telling them. For the first time, it came up with new hypotheses. But it went even further: It proposed processes to test the hypotheses, and data and sources to carry out the tests! And all so fast and for free! The most flabbergasting thing was that DeepSeek shows you its reasoning if you want, and it was spot on. It felt human, like a very intelligent and rational person. All the doubts the AI had during its reasoning were valid, and they helped me to improve my prompting tremendously. For example:
How did DeepSeek create such a good model for so cheap? It’s unclear, but it looks like it’s a combination of factors that I’ll explain in a premium article, where I’ll also cover the steps left to AGI.
What is salient here is that they didn’t require human feedback! They engineered a process where they simply fed data to the machine and structured ways for the machine to teach itself. Remember how I told you earlier that synthetic data would unbridle AI progress from data constraints? This is a perfect example.
This is very important because, in the past, the biggest bottleneck to AI progress was always humans. AlphaGo had to learn from tens of millions of human moves before it could beat humans, and that process took months. AlphaGo Zero, however, learned to beat humans without any human input, by playing against itself, in just 3 days.
AlphaGo Zero beat AlphaGo 100-0.
What does all this mean? It’s time to put it all together.
The best AI models were about as intelligent as rats four years ago, dogs three years ago, high school students two years ago, average undergrads a year ago, PhDs a few months ago, and now they’re better than human PhDs in their own field. Just project that into the future.
The makers of these models are all saying this is scaling faster than they expected, and they’re not using weird tricks. In fact they see opportunities for optimization everywhere. Intelligence just scales up. They believe AGI is 1-4 years away. This is only a bit more optimistic than the markets, which estimate it will arrive in ~2-6 years.
AIs are now like elite software developers. And AI is already improving AI. Put these two together, and you can imagine how quickly AI development speed will accelerate.
At this point, it doesn’t look like intelligence will be a barrier to AGI. Rather, it might be other factors like energy or computing power.
Except DeepSeek just proved that we can make models that consume orders of magnitude less money, energy, and compute.
So although electricity, data, and especially compute might be limiting factors to AI growth, we are constantly finding ways to make these models more efficient, eliminating these physical constraints.
In other words: AI is progressing ever faster, we have no clear barriers to hinder their progress, and those in the know believe that means we’ll see AGI in half a decade.
To answer the questions I posed in the intro: We’re jumping from the precipice right now, and odds are it will take a few years (1-6 depending on who you ask) for us to learn to fly, or to smash into the ground.
How can we prepare for that moment? This is what I’m going to explore in the coming articles:
What’s the latest on the impact to specific jobs?
Will AI accelerate science?
What have we learned about alignment? Will AIs kill us?
How can you get into AI, if you know nothing about it? How can you ride the wave?
I’ll explore these ideas in the upcoming articles!
An artificial intelligence that can do any task better than any human. Usually this refers to cognitive tasks, but it is believed that the jump from cognitive to real-world tasks won’t be tough. I’ll write about this in a future article.
Which he defines as expert-level science and engineering.
We went from o1 to o3 because OpenAI is terrible at branding, and because O2 is an actual, famous brand owned by Telefonica.
Here was its more detailed answer to my prompt, I assume because I have system settings that instruct ChatGPT to be detailed and break down problems:
Based on the information provided:
A and B are watching TV together.
B is eating chow mein while watching TV.
D is sleeping.
E is playing table tennis.
B leaves the room to answer a phone call.
Given that there are five people in the room and E is playing table tennis, they would need a partner to play with. Since D is sleeping, and A was watching TV with B (who just left), A is likely still watching TV. This leaves C as the only person who could be playing table tennis with E.
Answer: Playing table tennis with E.
Because most AI companies are losing money because they invest (in NVIDIA GPUs) to grow!
Sure, renewable electricity is not firmed (made stable across time), but battery costs are plummeting, and it looks like Trump’s presidency is unleashing fossil fuel production too. No matter how you slice this—whether through natural renewable capacity growth, new nuclear capacity, or fossil fuels, we will have enough energy.
Make sure you click on the DeepThink model to get the R1 model. I won’t recommend mobile app links because who knows what data they could be gathering about you.
It’s not clear, because they could be hiding how much compute they’re using, maybe to hide the fact that they circumvented the US ban on chip exports. But it looks like they did build a much cheaper model than their competitors.
2025-01-28 00:01:56
I’m pulling articles from future weeks into this week because the news are hot on several important topics. I might or might not publish additional articles in the coming week(s). The next batch is on AI and Tesla. Also, before we start, a poll for you to guide future articles:
Israel and Hamas have signed a ceasefire agreement. It’s supposed to have three phases that bleed into the reconstruction of Gaza over the next few years.
It won’t last that long. I don’t think it will make it to the end of the year.
Here’s what the agreement says and why the current geopolitics stamp an expiration date on it.
It has three phases:1
In the first 42 days, 33 Israeli hostages will be released (women, children, and older people)2. Israel will release hundreds of Palestinian prisoners—many of whom have life sentences for crimes against Israel. Israel will also start withdrawing from some parts of Gaza.
In the following 42 days, Hamas releases all the remaining hostages who are still alive. Israel releases more Palestinian prisoners. Israel should be withdrawing from all of Gaza.
After that, Hamas releases the remains of the deceased Israelis it holds, while Israel releases still more Palestinian prisoners. Israel and Hamas negotiate the future of Gaza. Hamas agrees not to rebuild its arsenal, and Gaza is rebuilt.
So why won’t this last?
For Israel, Hamas’ attack of October 7th 2023 was a holocaust. The Israel from before is dead: Israelis no longer believe that peaceful cohabitation with Hamas is possible. Now, the only thing Israelis think is how can we ensure that this doesn’t happen again? Never again.
This is why Israel’s Prime Minister Netanyahu has consistently said that Israel has three war objectives:
To return all hostages
To eliminate Hamas’s governing capabilities
To ensure that Gaza will no longer pose a threat to Israel
He reiterated this as late as January 18th 2025, after the ceasefire was signed!
So why is Israel signing the ceasefire? I don’t think broad international pressure mattered much: If anything, it has shrunk over time, as people have gotten used to the violence. It’s also not money: For all the war effort in 2024, Israel’s debt to GDP ratio has grown 8 percentage points, from 61% to 69%.3
There are two reasons why Israel signed the ceasefire. The most important is because the first war objective is to get the hostages, of course. They are the symbol of October 7th. If it can get them back, it can close that painful chapter.
But that is going to happen in the coming 84 days. It will be done by the end of April. After that, of course Israel will want the remains of the deceased, but that might drag out over time and is not as important as getting the live hostages.4 And with every new remains brought home, Israel will have less motivation to keep the ceasefire intact.
By April, Israel’s first war objective will have been achieved. It will then focus on the other two: the end of Hamas and the neutralization of Gaza as a threat. Do you really think Israel will let Hamas take over again?
Israel benefits from the ceasefire for a second reason: Trump. As Israel’s Prime Minister Netanyahu told Trump for his inauguration:
Trump wants to look like he can bring peace to the Middle East easily, so he pressured Israel. Israel knows that, when things fall apart, Trump will blame Hamas and have Israel’s back. And right now Netanyahu’s government is a bit more stable than it used to be. So he probably thought: OK let’s sign this deal, this will get us our first war goal and give us some points with Trump. Then, once we get the hostages, we can make the ceasefire fail, blame Hamas, and go for the other two war goals: Eliminate Hamas and make sure Gaza is never a threat again.
Why is the elimination of Hamas so important? Because Israel now realizes the group will never stop until Israel is eliminated. Why?
The updated Hamas charter sees peace with Israel as impossible, doesn’t give up the right to fight, doesn’t recognize Israel, and advocates for the liberation of all of Palestine—that is, Israel, Gaza, and the West Bank. Jerusalem must be the capital of Palestine. All Palestinian refugees would be entitled to both the right to return and reparations for the time they lost their homes.
Hamas claims it doesn’t have problems with Jews, but it can’t stand Zionism—the construction of a state for Jews in Palestine.
The popular opinion in Gaza still heavily supports Hamas: As of September 2024, 89% of Gazans thought Hamas didn’t commit the atrocities it committed against civilians, 40% thought Hamas was correct in attacking Israel, 39% thought Hamas was managing Gaza well, 36% wanted Hamas to stay in control of Gaza, 48% of voters would have voted for Hamas if there were elections, 36% still thought armed resistance against Israel was the way to get an independent state.
Hamas has built its entire identity around the conflict with Israel. That’s why it has diverted immense sums of money from the economy into building its weapons and tunnels. Have you heard any reference to a change of heart?
Palestinian authorities don’t recognize Israel’s existence,5 consistently portray Palestine as including Israel, raise children to believe all of Palestine should be free from the river to the sea,6 educate them to kill Jews, convert them into terrorists, funds the families of martyrs, and often quote any concession from Israel as a stepping stone to liberating all of Israel and Palestine.
Before, Israel could bury its head in the sand and avoid this reality, but then Hamas yanked it out and raped it. And nothing Israelis see make them think Hamas or Palestinians have had a complete change of heart. This is the new head of Hamas:
This is Hamas with Gazans:
This is what Israel sees, so it concludes that the only solution to Hamas is its total destruction. And Israel also believes that now is the perfect time to eliminate Hamas, because it is at its weakest.
I don’t think Hamas realizes how clueless its miscalculation was. Hamas thought:
They were in the right to kill, torture, rape, and kidnap thousands of Israelis
The Israeli retaliation would unify Palestinians and rally them to the cause
It would fire up its Muslim brethren, who would support Palestinians
In particular, Iran’s proxies, Hezbollah, Syria, and maybe even the Houthis, would come to their rescue
Israel would become an international pariah
The US would intervene to stop Israel’s attacks
Allah would support them in their Jihad
Eventually, all this friendship for Hamas and hatred towards Israel would result in a huge victory for Hamas and the recognition of the State of Palestine
Something something Hamas conquers back Israel, from the river (Jordan) to the sea (Mediterranean)
What happened instead? Hamas is alone.
The worst for Hamas is what happened at home. Although a sizable share of the population still supports them, that number dropped precipitously in the last few months:
As the war raged on, support for Hamas waned.
Gazans think Hamas did a poor job.
We can get a visceral sense of that—While Hamas tried to convey a sense of power during the first hostage release:
This was the reality:
So the internal support has fallen, but the external support has plummeted even faster!
They’re off the gameboard.
Lebanon and Syria are done. In Lebanon, Hezbollah’s leadership and thousands of its operators have been wiped off the map. The Lebanese are fed up with Palestinians and how they dragged them into their civil war and constant conflict, and how their politics have been controlled by the Iranian proxy rather than local politicians.
In Syria, Assad is out of the picture. The locals are fed up with war, they’re fed up with the hostility that Palestinians cause in the region. They seem to want good relationships with Israel, now that Syria has no military power whatsoever and is embroiled in its own conflicts with Turkey and Kurds.
The Houthis are in Yemen, too far away. They played at antagonizing Israel by launching missiles, but Israel bombed their ports, airport, and military facilities. They are in no position to seriously challenge Israel.
Iran has lost its proxies everywhere, has economic problems, its attack and defense systems have been exposed to Israel’s, it will be crushed economically by Trump, and it has rampant internal dissent. It will be hard for it to have much impact in Palestine in the coming years.
Arab states don’t support Hamas, or even Palestinians.
Obviously, that’s not what they say. They all show solidarity with Palestinians and their cause, blame Israel, and demand de-escalation. But the reality is different.
When Iran sent missiles to Israel, Arab countries sided with Israel. Jordan and Saudi Arabia actively helped Israel and the US down the missiles. They didn’t just let Israel and the US use their airspace: Jordan actively shot down missiles, too! Other countries, like the UAE, also shared intelligence with Israel and the US. Radars in Qatar helped identify the incoming missiles.
Egypt and Jordan have been at peace with Israel for decades. Egypt has had its border with Gaza closed for ages, and forced all its border population to move away from the border to create a buffer zone that tunnels couldn’t bridge. Jordan expelled the Palestinian leadership, went to war with it, and has serious concerns about the Palestinians living in Jordan. The government there collaborates heavily with Israel, including getting a lot of water from it.
The United Arab Emirates, Bahrain, Morocco, and Sudan signed the Abraham Accords with Israel—peace treaties—and didn’t back out of them.
Whereas Biden’s administration was sympathetic to Palestinians, Trump’s administration is radically more pro-Israel. Trump has described himself as the "most pro-Israel President", has said he’s done the most for the country, has told Israel to “finish the problem”, and has criticized Biden for limiting arms sales to Israel. Trump’s nominee for Secretary of Defense recently said:
I support Israel destroying and killing every last member of Hamas.
Also:
Why the ceasefire then?
Israel has three war goals: Get the hostages back, eliminate Hamas, and make sure Gaza is not a threat ever again.
The ceasefire allows Israel to achieve Goal 1.
It also wins points with Trump, hence the timing of the agreement (and not when Biden proposed it eight months ago).
Trump’s strong support of Israel allows Israel to break the ceasefire without repercussions from the US.
The moment Israel gets the hostages, it will focus on Goal 2: Eliminate Hamas, because Hamas still fundamentally wants the destruction of Israel.
Hamas is at its weakest, both internally and externally. It has never been easier to eradicate it.
Meanwhile, Israel is extremely strong, on the back of successes against Hamas, Hezbollah, Assad, and Iran. So Israel won’t let the chance pass.
It has an incentive to attack Hamas as soon as possible, to avoid its reinforcement.
All this tells us that Israel will break the ceasefire at some point after the end of Phase 2, when all the live hostages are in its hands.7 The end of Phase 2 is April 13th.
Meanwhile, it’s very unlikely that Hamas will break the ceasefire:
It’s super weak
It knows it will never get a better deal than what it got under President Biden
From all this, we can conclude that the ceasefire is most likely to be broken soon after April 13th. Do you agree? You can make your predictions here:
But wait, we’ve only discussed Israel’s war Goals 1 and 2. What about Goal 3, ensuring that Gaza is never again a threat?
2025-01-24 21:02:56
This is the 9th and final article of the series Where to Create Ten New Cities in the US.
The groundwork for this series came here:
The first eight articles in this series are:
So here are the ten cities I propose, along with takeaways.
Transform the military base into a Hong Kong of the Caribbean:
Boca Chica is the ideal point in the US to serve as the trading hub between the Earth and Space.
It would be a new city in the middle of San Francisco, unburdened by what has been—because it’s federal land, so it could be exempt from local and state regulations. This liberation would make it the fastest source of wealth creation in the history of humanity.
It could have been an amazing city if the Salton Sea hadn’t run out of water. We could revive it with more water, and make it into a walkable, futuristic city. The water would come from the Pacific, as the Salton Sea is below sea level.
It was a summer destination when the Salton Sea had water. Give it its water back, and push its development as the summer destination it should always have been.
They were both deprived of water because of LA’s thirst. With desalination, we can give LA the water it needs and reserve these lakes’ water, so they can grow back into communities. Owens Lake will likely remain a small agricultural community, but Mono Lake could evolve into a new Tahoe, except on drier land. Walker City could become an outcrop of the nearby military base.
It’s an existing lake between Lake Havasu City and Las Vegas’ Lake Mead. It has barely any life in it. It’s all federal land, dedicated to human enjoyment. This enjoyment can be improved, many jobs created, and many great lives lived, if we incorporate it into a city.
There are dozens of valleys between California and Colorado that could be filled with new lakes and host new lives, from Dixie Valley in Nevada, to the Escalante Desert or Sevier Lake in Utah, to Green River in Colorado. Each new lake could host a new, different, original city.
These are new communities near existing cities. This is the most common type of new community development, which depends in a big way on local regulation. The key for these communities is to sell a grand vision, but share it with the right people only at the right time.
There are dozens of such projects in the US at any given time.
It’s not easy to build new cities. Most sites have been tried already; if they didn’t work, it’s because there was something wrong about them. The result is that new cities must have a strong argument going for them. They’re either
Some exceptional situation, like Guantanamo Bay
Otherwise poor sites that become important due to some new technology or geopolitics, like Starbase
Sites that are missing a key ingredient—usually water—but can either regain it or get it from scratch
Sites close to existing and successful cities
Sites that are less than ideal, but are viable and could beat better sites because they have become complacent—especially through regulation
In other words, new cities are usually water arbitrage, regulatory arbitrage, extensions of existing cities, or simply unique spots.
2025-01-23 21:02:14
This is the 8th short article of the series Where to Create Ten New Cities in the US.
The groundwork for this series came here:
The first seven articles in this series are:
You don’t need to read them in order. Enjoy!
Companies used to build new cities:
Many new cities have been built in the United States since World War II, including Irvine, California; Reston, Virginia; and Walt Disney World. Irvine, established by the Irvine Company and incorporated in 1970, has grown to a population of 300,000. Reston, inspired by the Garden City movement, now has more than 60,000 residents. Disney World secured substantial legal concessions from Florida, which functionally delegated county governing authority to it, though the state legislature recently revoked those concessions.—Mark Lutter and Nick Allen, Building Freedom Cities.
The best place for new cities is as satellites of existing ones, because they can tap into local demand.
Can companies build new satellite cities? Yes, that’s what companies like California Forever are famously doing. But they’ve been mired in controversy: secret land buy, lawsuits, cantankerous media, ballot wars… Will they be able to prevail over all this conflict? I wanted to know, so I talked with the founder of California Forever, Jan Sramek. I also talked with a couple of developers trying to build other cities, one in Utah and another one in the Texas Triangle.1 Here’s what’s going on, and how to build new cities.
California is mostly empty. The moment you leave urban areas, you encounter emptiness.
You can see this in the distribution of people (red) in the San Francisco Bay Area:
This is some of the most valuable land on Earth!—one reason why my third city proposal, Presidio, aimed to build a new city here. So any land immediately adjacent to what is already built up would have tremendous potential. Especially here:
This is mostly Solano County, and it’s quite empty. That made sense when SF and Sacramento were two distinct urban areas, but now, given remote work and the cost of living, people are willing to live in one and work in the other. I met so many people who live in Sacramento or Stockton and commute to San Francisco, Palo Alto, or Mountain View every day! That’s over 3h of transport per day! A community here would make so much sense, the same as Irvine made so much sense between Los Angeles and San Diego.
So this is what California Forever wants to do:
This is what it would look like:
Does this land contain forest we would need to cut? Natural reserves with endangered animals? Cozy little farms we would need to raze? None of the above. It’s non-prime grazing land, a politically correct label to say nothing much grows here besides some grasses.
The area is not on an earthquake line, it’s 20 miles away from the closest forest fire risk area, it’s between 20 ft and 80 ft above sea level so it’s resilient to sea level rise and tsunamis…
This is such a no-brainer for construction that when the Department of Commerce analyzed the Bay Area in 1958, it predicted that there would be a city here in 2020.
As a good European, the founder of California Forever, Jan, thought that housing quality in the Bay Area was not that great, and that there were few, if any, walkable neighborhoods. Unacceptable given the high cost of housing. He spent one year looking for ways to accelerate building in existing Bay Area cities, but he realized it would fall short of the housing California needed.2 Fun fact, despite all the YIMBY3 successes over the last few years, and the resulting bills passed in the California legislature, the building of new homes is down!4
But he also noticed on a fishing trip to Solano County that there was a bunch of land there that didn’t have much built on, or orchards, or even crops growing there. It was also very well placed, so eventually he raised money and started buying it up to convert into a bustling city.
Such a city would bring hundreds of thousands of jobs and residents to the area.
But before you build a city, you need the land. Except if you announce that you will buy land to build a city, all prices will go up. So you need to buy the land without people knowing there will be a city here…
Of course, the moment the rumors started circulating, land owners started asking for inordinate amounts of money. California Forever accused them of collusion to keep prices high.
Then, some organizations started opposing the project:
Frances Tinney, an attorney for the Center for Biological Diversity, which evaluates developers’ environmental impact reviews, said that California Forever is “inherently unsustainable,” because it would require building entirely new water, wastewater, power and transportation systems, paving over ecosystems, increasing water use and polluting the air with constant traffic.
Well, building entirely new water, wastewater, power and transportation systems, paving over ecosystems, increasing water use and polluting the air with constant traffic is the definition of cities. According to this logic, humans should have no cities.
The logic is not just flawed for this reason. It also rests on mistaken assumptions. Concentrating people in cities is actually more friendly to the environment, because it limits the footprint people have on land, and because they emit less pollution per person when they share resources. Put another way: The current structure of Solano County (where California Forever would be) is less environmentally friendly than California Forever would be because in the current structure, people are dispersed.
And of course, given the current discourse on the environment, California Forever has tried to address it with its plans of a water recycling system, parks, circular economy, and the like.
The other typical criticism is primarily economic:
The buildout alone could leave Solano at a $103.1 million fiscal deficit, according to a report commissioned by the Solano County Board of Supervisors.
California Forever believes they would add a $40-50 million surplus.5
The media loved hating on California Forever, because it represents three infinitely clickable controversies:
Progress
Cement
Tech billionaires funding the two above
The move to withdraw the measure comes a week after a report by Solano county stated that the proposed city would likely cost the county billions of dollars, create substantial financial deficits, reduce agricultural production, harm climate resilience and potentially threaten local water supplies.—The Guardian
So the NIMBYs and media launched a campaign against California Forever. Spirits got heated. Moods soured, opposition rose. Despite that, Jan believes a ballot to approve the project and rezone the area will be ready for 2026, and it will pass. Why?
2025-01-21 19:15:10
This is part 7 of the series of short articles Where to Create Ten New Cities in the US.
The groundwork for this series came here:
The first six articles in this series are:
You don’t need to read them in order.
In the previous post, we saw proposals to build new cities near endorheic lakes—those whose water doesn’t flow into the ocean. But is that a fantasy? Is it actually doable? Yes, this is not a pipe dream. That’s exactly what Salt Lake City is:
And Provo!
Look at this community that developers are building on the shores of Utah Lake, in Provo:
Looks cool, right? They can do it because they have a great lake nearby. Same thing with Silicon Slopes, an area just north of this that hopes to attract tech talent.
There are other such valuable lakeshore communities, like those on Lake Tahoe:
Lake Tahoe, between California and Nevada, is one of the most valuable tourist destinations in inland California. Real estate prices there are through the roof.
What if we created another Tahoe, but in the desert?
We could make it whatever we want:
Or better, a mix of Tahoe and Vegas:
A Tahoe – Vegas – Dubai:
A city that would prove that the US stopped creating new cities in new sites simply because we ran out of will, and that with renewed will, we could settle it again.
Where would we put this city (or cities)?
Most lakes already host existing communities, and the ones I shared in the previous article are probably hard to turn into Dubai. So we would need to create new cities from scratch, which would mean creating new lakes. For that we have three requirements:
Lots of water
A flat area surrounded by mountains to contain the water
Most of that land should be available—that is, not taken already
Let’s start with the water.
Today, most of the Colorado River water goes to California:
But since California has a lot of coast, all of its water could come from desalination. We gave an example the other day with the revival of the Salton Sea.
What if we replaced all of California’s water with desalinated sea water? How much freshwater would we have to spare? What if we took some of that and sent it upstream, to Nevada, instead?
In the early 20th century, the Colorado River was diverted by mistake into the Salton basin, which was dry at the time, and the Salton Sea formed in just two years. That’s how much water the Colorado River carries. Similarly, within a few years, we could create a new lake in Nevada. Why Nevada?
Because a huge amount of land there is federal!
But not all federal land is the same. Some belongs to Native American reservations, some to National Parks. Surprisingly though, the majority is controlled by the Bureau of Land Management, so it doesn’t have much use!1
If we zoom in on the Southwest:
More than 85% of Nevada is federal land. Most of it is not National Parks, but is controlled by the Bureau of Land Management (BLM). Nevada has the highest concentration of BLM land!
There’s a reason for this, though: It’s pretty mountainous and desertic.
Las Vegas is in the middle of the desert, but it was made possible thanks to waters from the Colorado River. Could we do the same in some other part of Nevada? It’s worth looking at the land nearby. In this video, you first see Las Vegas, and then we make a trip inland.
Notice that some areas are green? These are usually the mountain tops that capture water from the air. The gray areas in between are the endorheic basins between them—like the Salton Sea, Lake Tahoe, or Mono Lake. You can see that in this map:
These are Nevada’s rivers. At the bottom right, you can see the Colorado River in blue. Do you see that most rivers here don’t go anywhere? That’s right, Nevada is mostly made of endorheic basins! And just a few thousands of years ago, most of these basins used to be lakes!
Any endorheic basin near the Colorado River would work. Here are some examples of such endorheic basins, with the Colorado below them:
2025-01-19 21:03:02
This is part 6 of the short series Where to Create Ten New Cities in the US.
The groundwork for this series came in these two articles:
The first five articles in this series are:
