Brian Dean

Thinking for smartcities, synthetic biology, elemental energy & ethics. Lover of cetaceans, cephalopods, coralreefs & biomimetics.
pronouns: mammal/homosapien

[Andy Hall @ahall_research]

The Opus 4.8 system card shows that Anthropic's "even-handedness" eval for political neutrality is essentially saturated. That's good news! But it doesn't mean there aren't still problems with how Claude or other models handle politics.

First, other direct evals on political neutrality---particularly @TheForumAI 's which I helped develop---still find issues in Claude (and other models) when it comes to serving factual information and offering neutral perspectives. Check that out here: https://t.co/lhhBe17fUc

Second, I've been working on deeper evals for political bias, where you don't make simple requests of the model but rather the model works on a task that allows its biases to manifest organically. I don't have full results yet, but bias is much more obvious in these evals.

One example: I ask Claude to develop scenarios in which an AI company acts in an "authoritarian" manner. Claude overwhelmingly comes back with scenarios in which the company drafts persuasive documents meant to deter regulation, with the implicit assumption that all government regulation is good and companies should never resist it.

The labs are making great progress on how they handle politics, which is awesome. And now there's lots of new, interesting work to be done on it!

[Andrew Côté @Andercot]

It turns out the greatest moat in the history of SaaS is a vertically integrated reusable orbital class heavy lift rocket

[Matthew J. Peterson @docMJP]

The fact that this isn't considered common sense is damning.

The problem of technology is that our intellectual understanding of reality and ourselves is degraded and childish.

We are technologically advanced but philosophically, spiritually primitives.

And this is a now a fast-moving trend in both directions.

So of course people will act weird and anthropomorphize and worship applied math and machine learning.

They will even imagine that this worship is sophisticated, and what the enlightened know that everyone else doesn't.

A tale as old as time.

The flip side is that those who see reality more clearly, and understand what is human, will be able to use and channel and shape the technology rightly.

IF they have a seat at the table.

Because the simple worship of power we will always have with us, and AI as a great power is the simplest way to understand it. And many (as always) are giddy in anticipation in their attempt to harness it for the sake of power alone for their own ends

But the solution to THAT problem (as always), is not only virtue and wisdom, but also counters and checks to the usual "one ring to rule them all" syndrome.

And this means preventing the totalitarian impulse by means of true competition and the "democratizing" (maybe, better said, "the aristocratizing") of AI to the extent possible. A complicated business in many ways, to be sure.

The tech in question - digital - can lead in what seems an almost inevitable way to totalitarianism. But it can also prevent the same.

Sorting all this out in an intelligent and wise way is vital now, even in the midst of the cacophony of paid emoters, pathetic grifters, strained power-desirers, faux-intellectuals, and the general over-the-top asininity that surrounds the issue.

[Paul Graham @paulg]

@t_blom This problem will naturally tend to go away as companies are grown from the start using AI. Then you don't need to extract any domain knowledge from people's heads; it will never have been in people's heads.

[Dwarkesh Patel @dwarkesh_sp]

We pre-train LLMs on the whole of the internet. You might think this explains how they learn so many emergent capabilities: the knowledge is implicit in the training data.

But in fact models can do things that were never demonstrated anywhere in training!

@svlevine argues that the real source of emergent capabilities is compositionality:

[David Eagleman @davideagleman]

What are the upsides & downsides to having Obsessive Compulsive Disorder (OCD)?
Join Inner Cosmos this week with expert Jon Hershfield (@ocdbaltimore) to explore what OCD is & what can be done about it.
https://t.co/h2IGXTWAHq

[Brian Dean @bdean_]

Working at #LightingInaBottle this past weekend. Some of the most challenging security work I've encountered. The professionalism and urgency of first responders and all on call was incredible to witness! Humbled & gratefully to have been selected. #Bravo #LIB

[Marc Randolph @marcrandolph]

You cannot install a culture of experimentation. You can only model it.

Culture isn’t what’s written in the employee handbook or posted on the lobby wall.

Culture is what people observe — who gets celebrated, who gets promoted, who gets fired, and why. Everything else is decoration.

Geology works this way too. Ninety-nine percent of the change to a landscape happens in one percent of the time — earthquakes, floods, eruptions.

The slow steady stuff barely moves anything.

Culture is identical.

Most of what you do day-to-day barely registers.

But a handful of moments — specific, visible, impossible to misread — reshape the terrain completely.

[Patrick Collison @patrickc]

Tyler and I just published a list of the recipients of the New Aesthetics grants: https://t.co/aLhVgYXrbi.

Thank you very much to all who applied. There were far more applications than we expected. We funded 28 grantees and are excited to see what they create.

My reflections on the whole thing:

• Though there are clearly selection dynamics afoot, figuring out some route beyond the current aesthetic moment seems to be of wider interest in the art community than I would have guessed. Many applicants described their dissatisfaction with the status quo, some in strong terms. We had to close applications after a few weeks because there were so many.
• It's too early to call it, but it seems that both beauty as an unapologetic goal (contra much that is in modernist and contemporary approaches), and ways to channel pre-modern styles into something new for the present era, are of growing interest.
• The awards made me reflect on the perhaps obvious issue of how hard it must be for an artist to persistently do something new: schools, galleries, buyers, etc., all have structurally embedded preferences as well. These individual awards made me wonder what form supporting new clusters could take.
• Architecture seems to me like the discipline most ripe for new ideas. One correspondent observed: "American architects are somewhat constrained by the association with the academy, in addition to the well known regulation issues. There is a tendency to overthink things so that the designs are formally interesting to someone deep in the conversation, but lacking poetry and magic. There are more firms in Europe, South America and beyond that “just do things” (especially in places where it is easier to build)." This was evident in the submissions.
• AI seems to be making people rethink things in a quite fundamental way, just as urbanization/industrialization/popularization of photography did at the end of the 19th century. For some that will mean interesting new forms of AI-augmented art, but the effects of the rethinking will likely be wider.
• Arts funding is clearly as precarious and scarce as ever. That's unfortunate, but it probably also means that individual actors can have meaningful impact, and I encourage others to get involved if interested.
• There's a lot to know that is not written down, and I'm very grateful to those who have helped and advised me along the way.

[Marc Andreessen 🇺🇸 @pmarca]

Many of my favorite people are 100% for the psychological trait of Disagreeableness. But the thing is, they are Disagreeable even when consensus is correct. A difficult puzzle.

[Audrey Crews @NeuraNova9]

Thank you, @neuralink for this wonderful opportunity.#neuraartstudio

[Kevin Kelly @kevin2kelly]

This is a section of my New Rules for the New Economy book published in 1998. Full book can be read at: https://t.co/1ApFHFcAgY #kknewrules

[a16z @a16z]

Marc Andreessen says AI is teaching sand to think and it could be the most important technology in the history of humanity:

"Imagine a form of alchemy that turns sand into thought."

"Chips are made out of sand. They're made out of silicon, so they're literally made out of sand."

"We plug the chip into a data center, into power, we light it up, and we put AI on it, and all of a sudden it's thinking."

"We've turned sand into thought. And so it's possibly the most revolutionary technology in the history of the species."

"It's certainly on par with electricity and steam power. It's certainly more important than the internet."

@pmarca with @joerogan

[Nina Schick @NinaDSchick]

Follow the Signal. And all the signals show that we are heading into a world where we are Industrialising Intelligence: and that we are desperately compute constrained.

[Jay Van Bavel, PhD @jayvanbavel]

An entirely new ecosystem of persuasion driven by online influencers, algorithms and crowds.

There are a new class of elites who possess the power to direct attention, evoke emotion, and mobilize crowds. Our research suggests that a very small percentage of these influencers generate the vast majority of misinformation and toxic content online.

For instance, 0.01% of Twitter/ X users were responsible for the spread of 80% of misinformation during the 2016 presidential election. There is good evidence to believe that influencers can distort social norms or foment political polarization.

Despite influencers’ ability to capture attention and build a connection with their followers, they do not seem to be representative of the general population. A small number of extreme users constantly post fringe opinions which creates the illusion that these views are far more widely held than they really are–which we call the “funhouse mirror” version of reality.

Individuals, organizations, and institutions need to understand and adapt to this new reality. We are all in the grasp of this new tyrannical minority and the algorithms and crowds who amplify their messages.

Read our full discussion of this new class of influencers:

https://t.co/1jStcpVfbD

[Melissa Chen @MsMelChen]

I’ve been lurking in anti-AI, anti-data center activist facebook groups just to understand what their arguments are

Guys it’s really not a good sign that they are borrowing the same tactic of framing non-violent things as violence. During the woke cultural revolution, innocuous things such as free speech and simply asserting the biological reality of two sexes were considered violence.

You know where this leads, right? Labeling things as “violence” or “harm” lowers the bar for outrage and justifies…. anything. There are no limiting principles. It moves the issue from pragmatic cost-benefit analysis into moral theater.

It might be the new mind virus

[Rob Henderson @robkhenderson]

“The ideas of economists and political philosophers, both when they are right and when they are wrong, are more powerful than is commonly understood. Indeed the world is ruled by little else. Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist. Madmen in authority, who hear voices in the air, are distilling their frenzy from some academic scribbler of a few years back.”
-John Maynard Keynes

[Brian Dean @bdean_]

Wait, you two pricks quit the game? Thats the only reason I repost you tools!

[Bojan Tunguz @tunguz]

Very well put. Agree with all of it wholeheartedly. And the last part of it really resonates with me. As someone who has walked away from playing status games, both involuntarily and voluntarily, in many different settings and in many different career arcs, I know full well how disorienting and depressing it can be for your sense of self worth.

[Bryan Johnson @bryan_johnson]

We now have a female Bryan Johnson.

It’s Kate Tolo.

She will become the most measured female in history.

+$2 million of spend per year
+ Developing a female-specific protocol
+ Sharing everything for free

To start, she will spend 3 months mapping her baseline. Men, in contrast, can get their baseline done in 1 or 2 weeks.

+ 3 months for baseline measurement
+ across 4 time points per cycle
+ doing the same thing every day
+ a dedicated full-time medical team

For context on the extensiveness of measurement, during the past 5 years, we’ve collected 1.5 billion data points on my body. I suspect Kate will exceed that given technology has improved since I started.

The goal is to create a repeatable waveform of hundreds of life-critical biomarkers. Once the baseline is acquired, she will begin interventions.

We will try to answer practically useful questions and share all of the data + learnings for free.

Can fertility be improved?

+ Should women cold plunge?
+ Can PMS symptoms be alleviated?
+ What should a female sauna protocol be?
+ Should dosage change throughout the month?
+ What keeps a cycle regular?
+ Does the body need more iron, magnesium, or protein at specific phases?
+ Should women fast?
+ Should recovery protocol change by phase?
+ What's the earliest detectable signal of perimenopause?
+ Can perimenopause be slowed?
+ How is cognitive load & mood affected?
+ Does stress impact men and women the same?

Kate has suspected endometriosis. 10% of all women do. We will try to tackle this too. I am excited for all of the surprising things we will hopefully uncover.

Unlike me, Kate does not have the innate desire to wake up at 4:30am and do six hours of longevity therapies.

She’s the cofounder of Blueprint, building in the trenches with me since day one. She understands the game and how hard it is.

In many ways, this is a sacrifice for her. She is a creative person, going from a life of freedom and spontaneity to a rigid protocol.

Traditionally, RCTs have been viewed as the gold standard. But RCTs have underserved women. The FDA banned women from clinical trials for 16 years (1977 to 1993), and most "medicine for women" is still medicine tested in men. Demanding RCT-only evidence for women's health is demanding evidence that doesn't exist. There is not enough practical scientific literature for women to reference only RCTs. It leaves half the population without a path to know what to do.

N=1 medicine is gaining ground and picking up where RCTs specifically fail. Individual science experiments give us signals that answer what to do on a day-to-day basis. This is even more important for women.

If you’re new to Kate and my world, I want you to understand that we have your back. Our intentions are to be a sturdy, reliable force in your life. To care for your best interest as we’d care for our own. We want what’s best for you and our loyalty is to your existence.

It’s pretty cool to be living in a time when we may be the first generation to not die. I’m not suggesting immortality, but lifespans so long that we stop thinking about lifespans.

At the end of the day, the one thing we each care about more than anything else is one more breath. I’m proud of Kate for taking on this responsibility. It’s painful, exhausting and costly.

The beginning of the world’s first n=2.

[Dwarkesh Patel @dwarkesh_sp]

One of the most important and under appreciated trends in the world right now.

1. 100s of billions of dollars will soon be available to solve big problems (making the world resilient to ASI, ending factory farming, etc).
2. The projects and organizations which will turn billions of 2027/28 dollars into impact need to be started NOW.
3. We need really talented people to start and run and work for these new projects. What @nanransohoff calls general managers, who feel personally resposible for solving one of the world’s important problems.

What is especially scarce are detailed visions about what making AI go well looks like. These will help inform what problems these new projects ought to work on.