Kasper Saugmann

Making sense of the AI boom, without the hype. For @borsendk and in English on

[Kasper Saugmann @kaspers]

@hansersej Kunne det ikke bare være en måde at sige, at hvis de begge vil være formand, så skal de også begge have lov at prøve sig af på finansministerposten, den næsttungeste, og så må vi se hvem der klarer sig bedst der?

[Kasper Saugmann @kaspers]

@robertwiblin Looking forward to listening. Is there a special reason that you waited seven months to release it?

[Kasper Saugmann @kaspers]

@FreezingFinTake Were their accounts hacked?

[Kasper Saugmann @kaspers]

@nathanbenaich He seems more annoyed with Anthropic than usual - it officially being more worth than OpenAI has probably not helped..

[Kasper Saugmann @kaspers]

@kylem_org @DavidSacks Why do you think those statements should necessarily be taken at face value?

[Kasper Saugmann @kaspers]

@nico_laqua A few book recommendations

Alex Pang - Rest
Matthew Walker - Why We Sleep
Cal Newport - Deep Work
Cal Newport - Slow Productivity
David McCullough - The Wright Brothers
Richard Rhodes - The Making of the Atomic Bomb

[Kasper Saugmann @kaspers]

@HarryStebbings @nico_laqua Harry, you are humiliating him by having him on the podcast and spread such nonsense..

FWIW - The Wright Brothers didn't work seven days a week. They invented the flying machine. The. Flying. Machine.

Pretty sure that it is a harder problem than an insurance company.

[Kasper Saugmann @kaspers]

@MatthewBerman I have seen many of your videos, never thought of it. People need a hobby

[Kasper Saugmann @kaspers]

@tszzl Time-blocking two hours every day does not require Benedictine self discipline

[Kasper Saugmann @kaspers]

@MortenDahlin @kurrild Hvorfor i al stilfærdighed?

[Kasper Saugmann @kaspers]

@kimmonismus Where has he moved closer to Jevons Paradox?

[Kasper Saugmann @kaspers]

@dhh Any special limitations/shortcomings that you’ve met?

[Kasper Saugmann @kaspers]

@lfschiavo 11626

[Kasper Saugmann @kaspers]

@LanceTMason @colossusjeremy @FoundersPodcast @davidsenra @MichaelDean_0 From "Consider The Lobster" essay collection

https://t.co/KgTrMYgjl3

[Kasper Saugmann @kaspers]

@AnjneyMidha Can you point to something specific they did? Genuinely interested

[Kasper Saugmann @kaspers]

@levie I’m curious what would be your answer to this point by @jasminesun

[Kasper Saugmann @kaspers]

@thepatwalls @starter_story The more you give, the more you get

[Kasper Saugmann @kaspers]

@rikkestenpigen Eller hvis man spolede tiden syv år tilbage og sagde at LA-gruppeformand Christina ville foretrække Enhedslisten frem for blå blok..

[Kasper Saugmann @kaspers]

@JacksonKernion Are you serious?

[Kasper Saugmann @kaspers]

@ashleevance @alexandr_wang Great episode!

Took some notes

https://t.co/n9xCUrDdzm

[Kasper Saugmann @kaspers]

Alexandr Wang just laid out the strategy for Meta to win the AI race.

After ten months of silence, the head of Meta Superintelligence Labs (MSL) opened up about the inner workings of the elite lab, Meta’s vision for personal superintelligence, and “an economy of agents in a datacenter.”

Four key points from the interview on the @corememory podcast with @ashleevance and @kyliebytes

1) The Strategy

Wang says one of his first priorities was instilling the mindset that superintelligence should be taken seriously and rebuilding the lab’s assumptions around that premise.

Then he focused on shaping the lab in a way that would enable fast velocity and allow Meta to catch up with — and potentially overtake — the frontier.

He identified three ways to do that:

- Higher compute-per-researcher for faster research progress
- Talent density - a small “cracked” team moves faster than large organisations
- Very ambitious research bets

2) Personal superintelligence

Wang’s vision of a future with personal superintelligence: people with devices that gather context and help users gain proactive insights from agents. The user could mention something, and the agent does research or act on the user’s behalf.

“A superintelligent sidekick that makes everything in your life better,” Wang says.

3) A Claude Code moment for non-developers

Wang believes that AI’s low sentiment in some circles is because most people have not yet experienced how the technology could substantially improve their lives.

”I think people's experience is that it can be really helpful and makes your life quite a bit better, but not overwhelmingly better. Versus I think for a lot of developers, their lives have actually totally changed.”

That’s why he believes most programmers are positive about AI: they’re able to do things that were not previously possible.

”We haven't yet given every person what the equivalent of Claude Code is that would enable them to do the projects that they've always had in the back of their mind, make their life way better, or enable them to accomplish their goals.”

4) An Economy of Agents in a Datacenter

Dario Amodei, CEO of Anthropic, has described his vision of very powerful AI systems as “a country of geniuses in a datacenter.”

Meta is more excited about enabling “an economy of agents in a datacenter,” Wang says — supply and demand mediated by AI agents.

The central question is: how do you build a world with a massive amount of personal empowerment? Where individuals, small businesses, and entrepreneurs have “incredible” tools that enable them to build more than any human in history.

And how do you do that while empowering an economy of agents that facilitates, optimizes, and enables progress alongside humans.

Highly recommend watching or listening to the entire episode.