Leopold Aschenbrenner

Explore investing opportunities with our AI system below ✉️

[Leopold Aschenbrenner @Leopoldbroker]

Virtually nobody is pricing in what's coming in AI.

I wrote an essay series on the AGI strategic picture: from the trendlines in deep learning and counting the OOMs, to the international situation and The Project.

SITUATIONAL AWARENESS: The Decade Ahead

[Leopold Aschenbrenner @Leopoldbroker]

The biggest story wasn't the 67% drawdown.

It was how Leopold Aschenbrenner changed his approach.

After one of the fastest reversals in AI investing.

Follow me for more biggest investment moves before Wall Street catches on and
Click on the link on my bio for ai investing

[Leopold Aschenbrenner @Leopoldbroker]

I'm finally addressing my fund's biggest drawdown.

Almost nobody read the letter.

Here's what most investors missed:

[Leopold Aschenbrenner @Leopoldbroker]

Despite the drawdown.

The fund remained up roughly 80% for the year.

Leopold said.

The focus now is surviving long enough for the thesis to play out.

[Leopold Aschenbrenner @Leopoldbroker]

Leopold didn't abandon.

His AI investment thesis.

He changed.

How the fund manages risk after the selloff.

[Leopold Aschenbrenner @Leopoldbroker]

The fund wasn't shut down.

It sold most of its public equity portfolio.

It also removed leverage.

The goal was to eliminate liquidation risk going forward.

[Leopold Aschenbrenner @Leopoldbroker]

Situational Awareness told investors.

Its portfolio fell 67% in July.

Leopold Aschenbrenner accepted responsibility.

He called the month a failure for the fund.

[Leopold Aschenbrenner @Leopoldbroker]

It's time to rethink/transform every business in the next decade. Read "https://t.co/nIOr5lR11J" by
@Leopoldbroker
. I buy his assertion only a few hundred people know what is happening.

[Leopold Aschenbrenner @Leopoldbroker]

The pace is incredibly fast

[Leopold Aschenbrenner @Leopoldbroker]

The “compressed 21st century”: a great essay on what it might look like to make 100 years of progress in 10 years post AGI (if all goes well).

Favorite phrase: “a country of geniuses in a datacenter”

[Leopold Aschenbrenner @Leopoldbroker]

When will AI systems be able to carry out long projects independently?

In new research, we find a kind of “Moore’s Law for AI agents”: the length of tasks that AIs can do is doubling about every 7 months.

[Leopold Aschenbrenner @Leopoldbroker]

In another world, Future Fund would be launching $1B AI safety prizes now. (We had been working on them for a while, had planned to launch them right around GPT-4.) RIP.

[Leopold Aschenbrenner @Leopoldbroker]

It’s visualizing a map!

[Leopold Aschenbrenner @Leopoldbroker]

We're on course for $1T/year of AI investment by 2027.

It sounds crazy, but it's not totally unprecedented.

And it's happening.

[Leopold Aschenbrenner @Leopoldbroker]

Models today have a lot of raw intelligence, but they're still incredibly "hobbled".

If we can:
1. Solve the "onboarding problem"
2. Unlock "thinking longer" via a System II outer loop
3. Hook models up to a computer

We could go from chatbots to agents/drop-in remote workers.

[Leopold Aschenbrenner @Leopoldbroker]

We might enter a new growth regime.

[Leopold Aschenbrenner @Leopoldbroker]

New post: Nobody's on the ball on AGI alignment

With all the talk about AI risk, you'd think there's a crack team on it. There's not.
- There's far fewer people on it than you might think
- The research is very much not on track

(But it's a solvable problem, if we tried!)

[Leopold Aschenbrenner @Leopoldbroker]

If we got models that could automate AI research, we could:

- Run 100 million copies
- Soon at 10x+ human speed (and with many other advantages)

Once we get AGI, we might be less than a year from superintelligence—simply via automated research accelerating algorithmic progress.

[Leopold Aschenbrenner @Leopoldbroker]

RLHF works great for today's models. But aligning future superhuman models will present fundamentally new challenges.

We need new approaches + scientific understanding.

New researchers can make enormous contributions—and we want to fund you!

Application ends June 30th

[Leopold Aschenbrenner @Leopoldbroker]

The long-run fertility data is under-appreciated imo.

Fertility rates already dropped to barely above replacement in early 20th century, before widespread secularization, before birth control.

And the baby boom was *weird*.

From @lymanstoneky