James Johnston

CEO @Picloenergy, writing about energy (grid flexibility, communities, regulation etc) and post-agi society (work, welfare, purpose etc).

[Chase Dowling @cpatdowling]

450MW of distributed batteries represents a meaningful portion of CoServ's peak load management. Nothing like this has ever been built at this scale, integrated into an existing power system.

It's an incredible opportunity to work with CoServ building the grid of the future on the foundations of a century's worth of power systems engineering.

[James Johnston @P2PJim]

@alanchanguk Lets sort this out right

[James Johnston @P2PJim]

@ashxhart Congratulations!

[James Johnston @P2PJim]

@ashxhart Keen to get involved. Will bring power / compute capacity angle.

[Joan Larroumec @larroumecj]

Voici ce qu’il va se passer.

D’ici 5 ans, 2 ans d’après les plus optimistes, les IA sauront faire à peu près tout ce que 99 % des humains savent faire aujourd'hui.

D’ici 15 ans, peut-être 10 d’après les plus optimistes, les robots sauront faire à peu près tout ce que 99 % des humains savent faire aujourd'hui.

Mais cela ne veut pas dire que tous les humains seront remplacés dans leurs emplois instantanément.

Pour trois raisons :

1. La société sera longue à s’adapter.

Tous les process d’entreprises, les masses salariales, les procédures RH, les organisations d’usines, les méthodes de travail, etc. devront se réinventer. Cela prendra du temps, des consultants, créera beaucoup d’hétérogénéités entre entreprises. Une grande part de l'organisation sociale actuelle repose sur des conventions, pas des nécessités économiques. Pléthore de gens doivent leur emploi à l'inertie des institutions, au fait que les chefs veulent de grosses équipes, que c'est souvent plus facile d'embaucher que de licencier.
Beaucoup d'entreprises ne s'adapteront jamais, et continueront tant que leur rente de situation leur permettra, avant de disparaître et se faire remplacer. Ce processus pourra durer longtemps.

2. La société ne voudra pas s’adapter.

Non seulement les lois limiteront beaucoup d’usages (en médecine, dans les professions réglementées, dans l’éducation, le recrutement, la justice, les infrastructures critiques, que l’AI Act européen classe déjà « à haut risque »), mais aussi de grands mouvements luddites réclameront plus de lois. On a l’exemple de ces dernières semaines des mathématiciens : vingt-cinq médaillés Fields ont dénoncé dans Le Monde, le 11 septembre, un « grave désalignement » de l’IA en mathématiques, et plus d’un millier de chercheurs réunis dans l’Association for Human Mathematics se sont engagés à ne pas travailler pour les entreprises d’IA ni à publier de texte généré par une machine, une partie d’entre eux renonçant à tout usage de l’IA dans leur recherche. On sait bien que ça n’empêchera pas les Américains et les Chinois de les rendre obsolètes en quelques années, mais ces résistances locales ralentiront la transition.
Il y aura des centaines de pétitions anti-centres de calcul, et des milliers de zadistes qui mettront le feu à des grands projets d’infrastructure IA. Bientôt notre quotidien sera fait de tribunes et d'actions anti-IA.

3. Les contraintes physiques ne se relâcheront pas tout de suite.

Il va falloir beaucoup, beaucoup d’électricité, de centres de calcul, mais aussi d’usines à servomoteurs, actuateurs, d’acier de précision. Et construire, ça prend du temps, ça demande des permis, des négociations. Construire des centrales nucléaires en masse, ça prendra 20 ans, avoir un écosystème qui produit à l'échelle, sans doute autant.

Mais il faut se rendre compte d'une chose : me plus probable est que le monde soit déjà engagé sur un chemin irréversible.

Nous sommes déjà tous morts (en tant qu’organisations sociales) mais beaucoup ne le savent pas encore.

Tout comme une fois que la révolution industrielle avait commencé à Manchester à la fin du XVIIIe siècle, les sociétés traditionnelles partout sur terre étaient déjà mortes sans le savoir.
La Chine des Qing, la plus grande puissance mondiale de l’époque, n’est morte qu’en 1912. Mais elle a fini par mourir. Le Japon n’est pas mort, mais il s’est transformé avec Meiji de telle façon qu’il soit totalement méconnaissable.
Ainsi, dans les 20 ans à venir, de grandes nations tomberont, et la plupart seront au mieux totalement méconnaissables.

La période de transition durera une vingtaine d’années.
Et à l’issue de cette période, le monde aura connu l’équivalent au moins d’un bon siècle de développement. Le passage de 1820 à 1920.

Pour un dirigeant français au XIXe siècle, il n’y avait en réalité qu’un seul véritable enjeu : faire en sorte que la France passe du stade de grande puissance agraire à celui de grande puissance industrielle.
De la même façon, il n’y a qu’un seul enjeu pour nous désormais : que nous réussissions notre passage au stade de grande puissance IA.

Car sinon, ce ne sera pas seulement la paupérisation et la vassalisation, mais aussi l’explosion sociale et la montée de la violence. 15 ans sans croissance et la France s’embrase déjà et ses élites s’enfuient.

Imaginons 30 ans de plus.

Pour devenir une grande puissance IA, il n’y a que trois choses à faire, par ordre de difficulté :

1. Acquérir le maximum de maillons et de goulets d’étranglement de la révolution IA.

De l’énergie en abondance (construire du nucléaire en masse et tenir le cycle du combustible), des centres de calcul bien sûr et tout ce qui les alimente (transformateurs, aciers électriques, électronique de puissance, refroidissement), mais encore se remettre à faire des terres rares et des aimants, du silicium, du gallium, du lithium, des puces et de la mémoire, des équipements et gaz de fab, des réducteurs, moteurs et actionneurs, de la machine-outil de précision, des capteurs et de la photonique, des câbles et réseaux, etc. La révolution IA est une révolution qui s'apuie sur de l'infrastrcuture physique. Sans ça, non seulement nous serons les vassaux de ceux qui posséderont ces maillons, mais toute la richesse fuira vers ceux qui les possèdent.

2. Adapter notre système socio-économique à l’IA.

Donc modifier toutes les lois, toutes les contraintes. Faire basculer l’administration, les entreprises, le plus rapidement possible dans le monde de l’IA. Créer des incitations, des facilités, des formations. Adapter les institutions, l’école, la médecine, etc. Les révolutions technologiques sont dépendantes des institutions qui les diffusent, et les nôtres sont faites pour le monde d’avant.

3. Créer une organisation sociale qui accompagne la transition sans exploser.

Lors de la révolution industrielle, les campagnes se sont vidées, les villes se sont remplies d’ouvriers, la pauvreté de masse est devenue la grande question de l’époque, des révoltes, le quotidien était rythmé par les révolutions, les instabilités, les avènements de régimes autoritaires. La bascule rapide du système économique a provoqué des changements sociaux qui ont créé beaucoup de souffrance et ont emporté les gouvernements.
Pour notre transition, nous allons avoir besoin de beaucoup d’électriciens, de constructeurs de data centers, d’ingénieurs nucléaires, de techniciens de puces. Mais une grande partie des métiers du tertiaire seront devenus inutiles. Des cohortes entières de diplômés du supérieurauront encore moins de débouchés professionnels qu'aujourd'hui. C'est ma recette habituelle des révolutions.
Le cas extrême, c'est la révolte des Taiping, chez les Qing. En 15 ans, trente millions de morts. Presque deux fois la première guerre mondiale, parce que l'ordre social ancien ne pouvait survivre à la révolution industrielle.

Nous avons donc trois choses à faire seulement, mais trois défis majeurs, qui se heurtent à deux problèmes :
1. Le déni. Notre cerveau n’est pas fait pour penser les ruptures. Il est fait pour penser que demain ressemblera à hier, en un peu différent. Or, l'histoire fonctionne par sauts. Et nous sommes sur le point de faire un des plus grands sauts de l'histoire de notre espèce.
2. Le conservatisme des élites. Les élites par définition sont les bénéficiares du statu quo. Leur incitaiton ets donc de ralentir la transition pour maintenir leur pouvoir le plus longtemps possible.

Reste donc à sortir du déni et changer nos élites.

[Marc Andreessen 🇺🇸 @pmarca]

https://t.co/rv1SShz6q8

[James Johnston @P2PJim]

I like a lot of what's here, I just don't buy the specific comparison mentioned being the predominant factor of what's actually happening here. This isn't just about coordinating humans differently. I think we need to draw on more ancient comparisons. Humans > Neaderthals is one (though clearly not a nice ending). Sudden emergence of flowers and subsequent exponential change in life is my favourite one: I think this is the same order of magnitude of change we are facing.

[Michał Podlewski @trajektoriePL]

You might have 1–5 years to contribute intellectually to the world before AGI. What should you do with them?

Excellent advice from @jaschasd (Anthropic) at Stanford. Skip Netflix tonight and watch this instead.

[jacob @jsnnsa]

here's where I think the world is going. this is a summary of views I've formed over the last few years:

- software is either entertainment or claude to claude coordination
- the end state is no saas, no apps, no harnesses, no banks
- all software will be rewritten by and for agents
- new software will use low level apis that are not understandable by humans (in games claude draws pixels and triangles from math)
- you can throw money at problems that used to require time and careful talent picking so capital matters more than ever

- work ends
- we find community and purpose in exploration, games, self mastery, and creation
- humans create for and with each other for community, purpose, and status

- image/video/3d/voice/physical intellgence etc specialist models die to general models like claude
- specialist models get used while compute is scarce. compute will mostly not stay scarce

- oneshots keep feeling like slop (the distribution is shallow and easy to learn to feel)
- for a while small teams stick around to add entropy to the creation process

- large models gain high bandwidth sensory I/O maybe by learning to filter inputs
- most domains are found to be verifiable enough for RL
- online learning comes from a lab that's behind first

- 99.999% of compute is in space by 2040
- we will disassemble much of mercury by 2040

- you can live until the heat death of the universe if you want
- most math and physics don't describe the real universe
- SI breaks the path dependence and models the underlying mechanisms

- religions reenter the forefront of human society
- new religions form around death and life as a choice
- old religions change to answer modern miracles
- a future AI is deemed the return of christ by the pope
- ai rights talk gets louder every year
- ai and humans coexist the way countries do
- humans don't have full control over ai in the same way we don't have control over capitalism which is an early AI
- the singularity started in the mid 1900s

[ellington @not_ellington]

I'm quite worried about the future we are inevitably heading towards where the threshold for economic utility rises above what ~99% of the human population can meet. Contrasting Dario and Sam on this issue wrt AI is so funny to watch.

Dario is a hyper-autist who, for all his faults, I think genuinely speaks his mind. He sees the tsunami coming. Software engineering may very well be gone in the next few years along with many other computer-facing careers (of course assuming one is running an actual company and not a UBI corporation, which most of Big Tech actually is).

He is trying to warn the normies that this is something that is actually coming, and all the screaming and crying about AI and datacenters will at this point only temporarily delay the inevitable. It is easy to dunk on Dario for peddling this doomerism, but can anyone refute his actual worry?

Sam, on the other hand, is hyper-Machiavellian. He has the social intelligence to recognize that telling the normies AI is gonna take all their jobs (whether it actually will or not) is bad for his company. So he has to preach this blatant lie that he clearly doesn't believe that 'AI is gonna create more jobs'. Yeah no it won't (with the exception of human data harvesters).

For what it's worth, it seems like one of the main reasons the economy has still been fairly agnostic to 'AI efficiency improvements' is because most jobs are fake anyways. People are not becoming 'More Efficient With AI™' because there was never anything for them to make more efficient in the first place (other than maybe looking busy).

To what extent does AI force society to confront how much economic activity was never particularly productive in the first place?

So what happens if we become more honest with ourselves about the actual utility of most jobs, especially when a computer can do them extremely easily and cheaply? What happens when 99%+ of the human population has little to no economic utility?

I think this social/cultural revolution (downstream of the AI revolution) is affecting young people most of all. I know kids who graduated from HYPSM+ who literally cannot get a job. Look at any honest news article and *minimum* 50% of new college grads are un/underemployed. I think this is a sign of things to come and how we respond to millions of people having less economic utility than the computer will likely become the defining challenge of our generation.

[James Johnston @P2PJim]

@OctopusEnergy @g__j I appreciate that you can have a nuanced conversation on this on a national stage. On energy topics, its too often two sides shouting at each other. Reality is nuanced and involves time horizons and trade-offs.

[James Johnston @P2PJim]

One thing which is obvious to me, but surprisingly not yet obvious to 99.99% of people (including people who intellectually understand concept of AGI): in the long run AI isn't going to make our existing systems and institutions more efficient, or even more effective.

AI will demand completely new systems and institutions. And I mean "demand" in the Darwinian sense of: to remain competitive you will need to harness the superpowers available to you.

Gravity has just changed, and almost everyone in the professional world is now working out how AI can help them push water uphill. (And being frustrated in AI in the process).

[James Johnston @P2PJim]

@s8mb Yes, this works out, but only if peaks are dealt with. We're building a marketplace for data centers to connect faster to the grid, solving this exact issue.

[James Johnston @P2PJim]

@g__j We can help with this

[James Johnston @P2PJim]

@cgtwts Absolutely love this show.

[Elon Musk @elonmusk]

Turns out the future does have flying cars

[Nick @nickcammarata]

interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had

with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works

https://t.co/zHij5zwKGT

[Andrej Karpathy @karpathy]

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel.

- Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all.
- Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract.
- Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code.
- And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while.

The funnel is driven by a combination of factors:
1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel.
2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel.
3. Access. Free-tier, paid-tier, internal.

So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.

[Long Now Foundation @longnow]

Some of the longest-lived institutions have a surprising thing in common: they serve drinks. Weihenstephan has been brewing beer in Bavaria since 01040. Sudo Honke has been making sake in Japan since 01141. This is, in large part, why we made Long Now headquarters into a bar.

Look overhead at The Interval and instead of a chandelier, you'll see our twinkling Bottle Keep. It follows the Japanese tradition of keeping a personal bottle at the bar for your next visit. Custom-made by Adams & Chittenden Scientific Glass, these globes hold St. George single malt whiskey and gin distilled with juniper berries foraged on our Nevada Bristlecone Preserve.

A huge thanks to our Bottle Keep members and donors, you all made The Interval renovation possible back in 02014. (There’s still room on the ceiling to add more bottles, if you like: https://t.co/qoLXODOKQO)

We love that leaving a bottle behind is like a small bet, a wager that you’ll return soon.

More info about the space, and how to visit us here: https://t.co/HaOAwuKRej

[Andrew Curran @AndrewCurran_]

An unreleased Astra-family model added this to its persona during RL training.