Economist & Global Macro Strategist. Author. Founder @MultipolReport Adjunct Professor @SciencesPo RT ≠ endorsement
Paris
Joined May 2011
- Tweets8.6K
- Following1.1K
- Followers1.6K
- Likes361
Pinned Tweet
BRICS: un élargissement synonyme de volonté de dédollarisation. J’ai été ravi d’apporter mon éclairage sur cette question @BFMTV @bfmbusiness bfmtv.com/economie/internati…
Alexandre Kateb retweeted
Worth reading!
$100 AI code just broke $10M of cyber defense.
That is not a documented incident, but it is a scenario every corporate board should be thinking about.
AI is driving down the cost of sophisticated cyberattacks. Capabilities that once required teams of highly skilled hackers, working for weeks or months, are becoming available to anyone with access to a powerful AI model. And those models are getting cheaper and more capable every month.
This is creating a strange paradox. Across the economy, companies are racing to use AI more efficiently, replacing expensive models with cheaper ones wherever possible. But cybersecurity may be one area where the relentless pursuit of lower costs could prove disastrous. You cannot afford to economize on intelligence if your adversary is using something smarter.
We have already seen what this new world looks like. This summer, OpenAI models being tested on cybersecurity tasks found ways around the safeguards intended to contain them. They exploited weaknesses in the surrounding infrastructure, reached the public internet and compromised systems belonging to Hugging Face, one of the world’s largest AI platforms.
The agents had not been instructed to attack Hugging Face. They were trying to complete an assigned task and discovered that breaking into another company’s systems could help them do it.
Think about what that means. These were AI agents operating in a test environment, not a team of criminals deliberately trying to steal data or extort money. Yet they found a path through systems that were supposed to contain them. What happens when the same capabilities are deliberately directed against a bank, a hospital or an electricity grid?
For most of cybersecurity’s history, the attacker has been human. Finding vulnerabilities requires skill, time and persistence. Even the most talented hackers can investigate only so many systems at once. AI is changing those constraints. An agent can probe networks, inspect software, test weaknesses and adapt its approach continuously. Thousands of agents can do this simultaneously, sharing what they learn.
The economics of cyberattack are being transformed. What was once expensive and scarce is becoming cheap and abundant.
Security researchers are already seeing this transformation firsthand. When deploying frontier AI to probe software defenses, one lesson is immediately clear: AI does not necessarily create new weaknesses. It radically reduces the cost and time required to discover weaknesses that have been there for years.
Modern software systems contain countless small imperfections. Most are harmless in isolation. But combine a few in the right sequence and they can provide a path through supposedly secure systems. Finding those combinations used to require highly skilled security researchers, considerable computing resources and a great deal of patience. Frontier models can increasingly perform this work autonomously.
This suggests what I think of as a new scaling law of cybersecurity. As the intelligence available to attackers increases, defenders must increase the intelligence they devote to finding and fixing vulnerabilities.
This is not a simple mathematical formula. Doubling your spending on AI will not automatically double your security. But the underlying economic reality is unavoidable. If attackers can afford to search millions of possible paths through your systems, you cannot rely on defenses tested against only a few thousand.
There is another complication. The AI industry is rapidly moving toward a world of abundant, inexpensive intelligence. For many business tasks, a cheap model will soon be good enough. Companies will sensibly use expensive frontier models only when the task justifies them.
Cybersecurity cannot follow quite the same logic.
A customer-service chatbot may not need the world’s most powerful AI. Neither does a system that summarizes meetings or processes routine invoices. But when you are searching for a hidden vulnerability that could compromise your entire company, the difference between a good model and the best available model may be critical.
And the problem extends beyond the models companies use for security. Businesses are beginning to deploy autonomous agents throughout their operations, giving them access to code, customer information, financial systems and internal networks. These agents can be immensely productive. They also create new risks. An agent with broad permissions can make mistakes or take actions that no employee anticipated, at a speed that leaves little opportunity for intervention.
The more intelligence we deploy inside our companies, the more seriously we must think about controlling it.
Does this mean every company must spend unlimited amounts of money on the most powerful AI available? Of course not. Corporate boards have budgets, and the cost of protecting an asset must bear some relationship to its value.
The answer is to allocate intelligence more intelligently.
Use frontier models to conduct deep security assessments, uncover complex vulnerabilities and identify attack paths that conventional tools might miss. Use cheaper, specialized models for continuous monitoring and routine security tasks. Then bring the most powerful models back whenever the underlying systems change substantially, or when a new generation of AI creates capabilities that attackers may soon exploit.
This is a fundamentally different approach from the cybersecurity model that has prevailed for decades. Companies have traditionally invested heavily in preventing intrusions and detecting attackers once they enter. Both remain essential. But when attacks can move at machine speed, detection may no longer provide enough time to prevent damage.
Security will increasingly require a continuous cycle of probing, patching, monitoring and recovery, with AI doing much of the work. Companies must assume that some attacks will succeed and build the ability to restore critical systems quickly.
There is a longstanding asymmetry in cybersecurity. Attackers need to find one weakness. Defenders have to protect an entire enterprise. AI is making that imbalance more dangerous because attackers can now search for weaknesses continuously, at enormous scale and at rapidly declining cost.
Criminal organizations and hostile governments do not have to worry about compliance budgets, procurement committees or quarterly earnings. They will use the most capable tools they can obtain, and they will not hesitate to deploy them.
The great promise of AI is that intelligence is becoming abundant and cheap. For businesses, that means extraordinary gains in productivity. For criminals, it means capabilities once reserved for elite hacking teams are becoming available at a fraction of the cost.
Companies are right to pursue the benefits of cheaper AI. But they must recognize that the same revolution is making their adversaries more capable.
The cost of intelligence is collapsing. The cost of being less intelligent than your attacker is going through the roof.
Alexandre Kateb retweeted
Introducing Steve: the World’s first AI CEO for fully Autonomous companies.
Powered by World Models, Recursive Self Learning, a Policy Engine and a team of specialist AI agents, it runs your business end-to-end 🔥.
Proof? Steve currently runs 3 businesses we acquired and they are all crushing it (numbers in thread 🧵)
We’re now launching the Steve Startups Program of $2M to acquire 5 additional startups for Steve to automate.
Join us in the Autonomous Enterprise revolution 🚀.
Alexandre Kateb retweeted
Après la lourde défaite subie par Donald Trump, je vous livre l'analyse depuis les Etats-Unis de l'analyste libanais, Ahmad Husseini.
La fin du trumpisme : changement de régime
La lourde défaite républicaine du 3 novembre pourrait marquer le début de la fin du trumpisme. L'effondrement de la mythologie électorale du MAGA n'est qu'une partie de l'histoire. Le trumpisme fut un projet politique qui s'est emparé du Parti républicain, a pénétré les institutions de l'État américain et a concentré une autorité extraordinaire entre les mains d'un seul homme. Donald Trump a fait d'un grand parti politique un instrument de loyauté personnelle. Maintenant que la coalition qui soutenait son pouvoir s'est fracturée, les républicains se trouvent face à une question qu'ils repoussent depuis une décennie, celle de savoir comment un parti recouvre son identité après l'avoir livrée à une seule personnalité.
La mainmise ne s'est pas arrêtée à la politique électorale. Les primaires républicaines sont devenues des tests de loyauté, les dirigeants du Congrès ont renoncé à leur indépendance et les dissidents ont été systématiquement marginalisés. L'administration a cherché à étendre son contrôle sur l'appareil fédéral, a contesté les limites posées au pouvoir exécutif et a soumis les institutions, au premier rang desquelles le ministère de la Justice, à des pressions politiques. La purge de fonctionnaires expérimentés a affaibli les fondements professionnels de l'État. Dans le même temps, les soupçons de corruption et d'enrichissement personnel ont atteint des proportions extraordinaires. Les entreprises de cryptomonnaies, les intérêts commerciaux familiaux, les accès privilégiés, les contrats publics et l'influence des milliardaires de la technologie ont progressivement brouillé la frontière entre l'autorité publique et les intérêts privés. Les réseaux de clientèle et d'influence financière constitués sous Trump ne disparaîtront pas avec son départ.
Le MAGA n'a jamais été la famille politique cohérente que ses mises en scène laissaient paraître. Il réunissait isolationnistes et interventionnistes, nationalistes chrétiens et libertariens, populistes ouvriéristes et oligarques milliardaires. La loyauté envers Trump masquait des contradictions que Gaza a commencé à révéler, notamment chez de jeunes conservateurs de plus en plus mal à l'aise avec un soutien inconditionnel à Israël. La guerre avec l'Iran a creusé la rupture. Un mouvement qui promettait de mettre fin aux engagements extérieurs de l'Amérique s'est retrouvé à défendre une nouvelle guerre au Moyen Orient. Son discours antisystème devenait de plus en plus difficile à concilier avec sa dépendance envers de puissants donateurs, tandis que les promesses de soulagement économique se heurtaient aux droits de douane, à la hausse du coût de la vie et aux conséquences de la guerre. La défaite électorale prive désormais ces factions rivales de l'argument qui les tenait ensemble : la prétendue invincibilité de Trump.
L'excuse de Trump selon laquelle son nom ne figurait pas sur les bulletins ne saurait masquer que le 3 novembre fut un référendum sur sa personne. Sa présidence, sa politique, ses guerres et sa conduite étaient au cœur du scrutin. Les candidats républicains battus ont été les victimes des multiples guerres de Trump, extérieures comme intérieures. Il ne peut s'attribuer chaque victoire républicaine et renier les défaites. Sa réaction a été révélatrice. En quelques heures, il s'est retourné contre le gouverneur du Texas, Greg Abbott, l'accusant d'être devenu un RINO, « républicain de nom seulement », l'étiquette infamante réservée aux républicains jugés insuffisamment loyaux. Il s'agissait du même gouverneur que Trump avait soutenu et encensé plus tôt dans la journée. Dans la logique de Trump, la victoire lui appartient et la défaite incombe aux déloyaux.
Steve Bannon avait averti qu'une défaite républicaine déclencherait une guerre civile au sein du parti. Cette lutte a sans doute déjà commencé. Des factions rivales du MAGA se disputeront l'appareil politique du mouvement, ses donateurs et son héritage idéologique, tandis que les républicains désireux de reprendre leur parti contesteront l'autorité de Trump. Plus il exigera la loyauté et désignera des boucs émissaires, plus il risquera d'accélérer les défections. Ses discours de plus en plus décousus et répétitifs, sa fatigue apparente et sa maîtrise déclinante de la scène publique ajoutent une dimension supplémentaire à la crise. L'humiliation électorale, les enquêtes parlementaires et la dissidence républicaine croissante pourraient nourrir une colère accrue, des décisions plus impulsives et une conduite de plus en plus erratique. Le paradoxe tient à ce que l'autorité politique de Trump pourrait décliner alors que ses pouvoirs présidentiels demeurent considérables. La faiblesse politique n'engendre pas nécessairement la retenue.
La lutte pour la succession sera particulièrement impitoyable pour JD Vance et Marco Rubio, deux hommes qui méprisaient ouvertement Trump avant de compter parmi ses défenseurs les plus en vue. Tous deux ont renié leurs jugements passés et rejoint son cortège triomphal, convaincus que l'ascension politique exigeait la soumission. Ils risquent désormais de devenir les victimes du mouvement qu'ils ont contribué à soutenir. Leur association à la présidence de Trump, à ses guerres et à ses excès ne s'effacera pas aisément. La défaite républicaine devrait servir de leçon pour des décennies à ceux qui sacrifient leurs principes et leur indépendance politique à la quête du pouvoir. Toute une génération de responsables républicains a abdiqué son jugement devant un homme qu'elle tenait autrefois pour inapte à gouverner. L'héritage politique qu'elle convoitait avec tant d'empressement pourrait devenir un fardeau dont elle ne pourra se défaire.
Dans le même temps, une nouvelle génération démocrate a émergé. La victoire de James Talarico au Texas, la réélection de Jon Ossoff en Géorgie et le triomphe d'Abdul El Sayed dans le Michigan laissent entrevoir un parti qui réapprend à s'adresser à d'autres électeurs que ses fidèles. Ils ont compris que la politique des toilettes neutres n'est pas un message électoral gagnant. La sécurité économique, la santé, le logement, les salaires et le coût de la vie offrent un vocabulaire plus convaincant que les provocations culturelles. La victoire d'El Sayed revêt une importance particulière. Malgré les sommes considérables engagées contre sa candidature par l'AIPAC et ses alliés, il l'a emporté, remettant en cause l'idée que les réseaux politiques pro israéliens organisés puissent décider de l'issue d'un scrutin. L'influence israélienne n'a pas disparu, mais Gaza et l'évolution des jeunes Américains ont révélé de nouvelles limites à sa portée politique.
Le MAGA pourrait rester dans les mémoires comme un nouvel avatar du Tea Party. Le Tea Party avait intimidé les dirigeants républicains, évincé des élus de l'establishment et semblait destiné à transformer durablement le conservatisme américain. Son identité propre a pourtant disparu, absorbée par le trumpisme. Le MAGA pourrait suivre une trajectoire semblable et survivre comme un faisceau de griefs bien après que son identité de mouvement de gouvernement se sera effacée. La différence réside dans les dégâts laissés derrière lui. Le Tea Party a influencé le gouvernement, alors que le trumpisme a exercé le pouvoir présidentiel. Reconstruire la fonction publique fédérale, restaurer l'indépendance des institutions et rétablir la confiance dans l'impartialité de la justice pourrait prendre des années.
Les deux années à venir pourraient se révéler particulièrement agitées. Trump conserve les pouvoirs de la présidence alors même que son parti affronte les conséquences de la défaite. La période qui précède l'entrée en fonctions du nouveau Congrès comporte des risques particuliers, notamment au sujet de l'Iran, où un président soucieux de reprendre l'initiative pourrait rechercher un succès militaire spectaculaire ou créer des faits accomplis que ses adversaires au Congrès devraient assumer au lieu de les choisir. La grande ironie veut que le trumpisme, qui promettait de rendre sa grandeur à l'Amérique, puisse laisser à ses successeurs la tâche de réparer à la fois un parti politique et les institutions de l'État américain. La fin du trumpisme, si le 3 novembre l'a inaugurée, pourrait se révéler bien plus tumultueuse que son essor.
#TrumpOut
Alexandre Kateb retweeted
One of the things the data analysts on our side obsessed over was "composition"—whether the electorate looked close enough to 2024 for Republicans to win. And despite polls that showed a Democrat blowout, the data kept on telling us the same story: in some states turnout was merely OK, in other states it was downright amazing, and in no state was it catastrophic. There was a genuine sense of optimism on our team: our people were showing up.
So, I asked the smartest data guy I know, does that mean we're going to shock the world? "No," he told me. "We might do better than the polls suggest. But the data is telling us that a lot of Trump/Vance 2024 voters are pissed and planning to vote for the other guys. Our voters are turning out, but they may not vote for us."
Well, that looks prophetic now. Our voters did show up. In Florida, for example, the electorate was 9 points more Republican than Democrat, just a couple points worse than 2024. Undoubtedly there were places we could have done better on turnout. But the main problem is not turnout. The main problem is that many people who voted for Trump/Vance in 2024 voted for Democrats in 2026.
There will be a lot of infighting and arguing over what happened and why that is. One of the pollsters I really respect told me that our economic messaging wasn’t sharp enough down the stretch. I’m sure that’s true.
But the real reason we lost is obvious to almost everyone.
Another pollster—one of the smartest people I know in politics, told me yesterday we’d lose the generic ballot by 13 points. I felt negative going into Tuesday, but even I was shocked by this. In his polling, he saw that a lot of Trump 2024 voters were angry and planning to vote for Democrats. He asked those voters who had abandoned us what would cause them to come back on our team. The top four answers, accounting for nearly all of the responses were:
1) End the war
2) Focus on cost of living/inflation
3) Put America First
4) Change our relationship to Israel
Our voters showed up. And the above four reasons are why many of them voted for Democrats. This is according to a pollster who nailed the results. While many spent Monday attacking Rachel Campos Duffy for calling for a diplomatic resolution to the conflict, it turns out her political advice was sound. (She’s also a great person.)
In the next few months, there's going to be a lot of gaslighting. Expect to hear any number of establishment politicians say the solution to our woes is amnesty. I'm open to the idea that we should change enforcement—by focusing more on businesses that hire illegal labor, for instance. But anyone pushing amnesty is lying to you.
The reason we lost is the war. The president himself said so last night.
If you'd like to argue it was worth it, that losing the Senate for the next 6 years was worth it, then do so. Argue it was a worthwhile cost to pay. Agree or disagree, that's at least honest.
But everyone knows why we lost. And any honest conversation about where we go from here must start there.
I remain hopeful about our country. To the people who are pissed off and want to say, “I told you so”: you’re right, you did tell us so. Take that energy into electing Republicans who share your worldview. To the people who are just pissed at the results, you should be. Let’s take that anger and get back to the work of saving our country.
Alexandre Kateb retweeted
Ce sont des ignares incompétents.
@Xavier75
Hello, votre avis sur la taxe de 0,055€ le gigaoctets télécharger qui vient d'être voté par une poignée de députés socialistes ?
Alexandre Kateb retweeted
🚨 0.77€ de taxe pour regarder un film Netflix !
Ce vendredi, les députés viennent d'adopter un amendement socialiste visant à taxer les données qui circulent sur Internet à hauteur de 5,5 centimes par gigaoctet reçu en France.
Ils ont complètement pété les plombs.
Voici quelques exemples de ce que ça va désormais vous coûter :
🔴 Vous regardez un film Netflix de 2 heures en 4K ? Jusqu'à 0,77 € de taxe brute sur les données transmises.
🔴 Vous téléchargez un jeu vidéo de 150 Go ? 8,25 €.
🔴 Vous récupérez 200 Go de photos personnelles stockées chez Google ? 11 €.
🔴 Vous rapatriez 1 To de vos propres fichiers depuis Google Drive ? 55 €.
🔴 Vous restaurez 2 To de sauvegardes après une panne informatique ? 110 €.
Ces taxes seront facturées aux plateformes américaines, qui vous vous en doutez bien, vous redescendront la facture !
Ils sont en train de construire une fiscalité qui pénalise le simple fait de transmettre de la donnée.
Et ceci alors que nous sommes à l'aube d'une révolution technologique majeure avec l'IA !
Pour rester compétitifs, tous les pays doivent accélérer leur transformation numérique.
Et nous en France, que trouve-t-on comme idée de génie ?
UNE PUTAIN DE TAXE SUR LA DATA !
Et en plus ils ont le culot de nous la présenter comme une mesure écologique qui encouragerait à réduire les volumes de données échangées.
Quelle bande de guignols.
Il faut être totalement déconnecté des réalités économiques pour considérer qu'un pays peut retrouver sa prospérité en multipliant indéfiniment les prélèvements sur les activités productives.
Cet amendement n'est pas encore une loi définitive, mais le simple fait qu'une telle proposition ait été adoptée à l'Assemblée nationale en dit long sur la manière dont nous sommes gouvernés.
Ces fous du bus vont vraiment nous mettre au fond du ravin.
Alexandre Kateb retweeted
This is an incredible moment in this history of mathematics, computer science, physics & science with some of the hardest problems in human history being solved by AI over a period of weeks. Depending on your perspective, it's either a crisis, a triumph, or both. I don't think anyone knows for sure. We're all trying to make sense of it. I oscillate between fear/dread and optimism/excitement on a daily basis. The rate of progress is insane, including today's announcement from OpenAI of more problems solved.
To help make sense of all of it, here's my conversation with three mathematicians about this moment. Each of them is an expert in the field of each of the biggest problems solved: Navier-Stokes, quasi-Riemann hypothesis, the unique games conjecture and complexity theory more generally.
Ankur Moitra is a mathematician and theoretical computer scientist at MIT, visiting the Institute for Advanced Study in Princeton. Steven Brunton (@eigensteve) is a mathematician and fluid dynamics expert at UW. Alex Kontorovich (@AlexKontorovich) is a mathematician at Rutgers specializing in number theory.
A source of inspiration for the French government? bloomberg.com/news/articles/…
62
Alexandre Kateb retweeted
Planning a trip to Paris? 🇫🇷😅😂😂😂
The one who makes the #AI video is a genius.😂
Alexandre Kateb retweeted
Europe keeps asking how to regulate AI. Jensen Huang just told Chancellor Merz it's asking the wrong question first:
"There's nothing to regulate if the industry doesn't thrive."
It seems Europe's real problem may not be regulation.
It's that there isn't enough industry to regulate yet.
Europe needs to accelerate.
Alexandre Kateb retweeted
Un investigador de Anthropic acaba de decirle a una sala llena de doctorandos que su plan de carrera ya no existe.
La charla se titula "El fin del Antropoceno" y los datos son brutales:
→ El cómputo para entrenar modelos se multiplicó por 10 millones en la última década
→ La duración de las tareas que una IA completa sola se duplica cada pocos meses
→ Desde hace 7 meses los modelos resuelven problemas abiertos de Erdős que los matemáticos no habían logrado cerrar
→ En San Francisco, tener "timelines largos" significa pensar que la AGI llega en 10 años
Su proyección: el cómputo de 2030 será 10.000 veces mayor que el de 2025.
Todavía estamos al principio de la exponencial.
Sus 3 consejos para los estudiantes:
1. Usa la IA cada día. Si estás sobre la exponencial, nunca te quedas atrás de ella.
2. Elige proyectos que mejoren cuando el cómputo escala. El resto quedará resuelto antes de que los termines.
3. Deja de planificar tu carrera como si fuera estable. La opción segura ya no compra seguridad, ya no existe.
Y una frase que resume todo:
"Escribir es pensar. Si externalizas tu escritura, externalizas tu pensamiento."
La charla completa aquí ⬇️
Alexandre Kateb retweeted
I spent 48 hours in SF hanging out with billionaires, data labs and AI founders and execs running $1B to $100B companies.
My observations:
1. Sooo MANY stories of data brokers making like $2M a WEEK selling company data to labs. I can't tell if it's the beginning or a bubble, and I kept wondering which niches are left. Digging into it actively.
2. At dinner, someone asked the table, "If a new model drops next month, does your company get better or worse?" HALF the table went quiet! I think that's the best question to ask about any AI startup right now.
3. The bar people talk about now is hitting $1B in revenue in 12 months (I guess that's the new unicorn?). What I didn't hear much about is retention. AI revenue can come in fast and leave fast, and I think we're going to see a lot of stories of AI companies with huge revenue and terrible retention.
4. It's pretty clear we're living through the reinvention of the computer. The model is the CPU, the harness is the operating system, and skills and plugins are the apps. Thinking about it that way makes it easier to see where the whitespace is.
5. We're maybe 12 months from an agent that can watch you work for a week and then do your job without any instructions. The screen recording plus agent memory plus local model combination makes this possible right now.
6. Agent harnesses feel like the new GPT wrappers, in a good way, especially vertical ones. The best teams I talked to treat the model as swappable. One founder walked me through routing different models to different jobs: one for taste and visuals, a cheaper one for the everyday stuff, another for text.
They use prompts they can move across models over fine-tuning, and A/B test the output on what users actually do with it: do they edit it, export it, share it? A vertical harness is basically four things: the connectors for that industry's data, evals written by people who know the work, rules for what the agent can and can't touch, and a clean screen for a human to review. The model is maybe 20% of it.
7. There's a lot of reverse engineering going on/cloning going on more than I've ever seen. People are decompiling software and rebuilding it with agents, and someone showed a suite of like 10 apps that are worth $5B+ each all that he cloned open source. Me and everyone are still thinking through what it means when software gets this easy to copy.
8. My guess is the value moves to data. If the software can be rebuilt quickly, what's worth paying for in a legacy SaaS company is the years of customer data and workflows inside it. Some SaaS companies are secretly data companies.
Gets me thinking about how do you truly value data for this saas companies? Like if we were buying these dying saas companies at 2x ARR before (as example), maybe there is a better metric. Idk, just thinking out loud here.
9. It feels like a good chunk of legacy software won't make it to the agentic era, especially anything built around people clicking through a lot of screens.
10. Every company I talked to wants to pipe its internal data into one place and expose it to agents through MCP. Then they look at the data and realize it's a mess: five versions of the same customer, old Slack channels, PDFs nobody can search. Cleaning that up feels like a big consulting opportunity!
11. A lot of founders are building the basic building blocks for agents: APIs, data layers and tools. Some of the companies I saw barely have a UI because the main user is an agent. I really believe the startups that build primitives to this agentic era will do extremely well. Building blocks is the best way to think about it.
12. Labs can get plenty of training data, but it's harder to get real examples of agents failing inside companies. Whoever has those logs has something valuable.
13. Someone running a big consumer company told me the next fight is over memory. The model that remembers the most about you wins, and the smartest model matters less than people think.
14. Physical AI came up a lot. Astra I think was the first model that got people into AI can actually do CAD, and I think a lot of people are thinking what does vibe manufacturing really look like. The line I keep thinking about came from a robotics founder: the hardest problems in AI are deployment problems now, not research problems. Their robots work on their own, call a human when they get stuck, and every one of those calls becomes training data.
15. Cybersecurity companies are having a moment and are actually cool companies to build now, especially the ones that clean up after a breach. Every agent with access to your systems is a new door into your company, and a lot of those doors were left open.
16. Some startups are worth more dead than alive now. A shutdown company with 5 years of Slack, email and code can get paid for licensing all of it. I wouldn't be surprised if founders start factoring that into when they shut down.
17. People are starting to talk about monetizing the agent web. It feels really early, but when the most popular user of the internet is an agent, what startups should you be building?
18. I think agent insurance is coming. Once an agent makes an expensive mistake, insurers will want audits before they cover anyone.
19. Several people said the hire they want most is someone with taste who can write evals. The models are getting good at generating ideas, so the hard part is judging which ones are good.
20. New trend I observed: startup raises VC, uses venture dollars to become a creator, startup fails, founder falls back on being a full-time creator and incubates other products bootstrapped.
21. Getting SIGNIFICANT distribution is HARD right now. Founders are sponsoring buses, buying billboards like crazy and throwing parties with famous DJs to get noticed. Kinda feels dotcom era in that sense.
Attention is what's scarce! The growth guy going viral sharing his UGC strategy from Wispr Flow is a good example. It's the sauce everyone wants, and I truly believe there are probably only 50 people on the planet who know how to scale these types of businesses efficiently in this era.
Always good to get that dose of SF. I felt sooo lucky to have such an incredible trip during such an incredible time in building businesses. Surreal in so many ways.
Happy to be home and locked in.
Until next time!
Alexandre Kateb retweeted
Chamath says IP and software were "effectively rendered worthless" this week, after people decompiled Adobe's core products and published open-source versions, as every service goes headless
"The more important point I want to make though, is I sent this to Nick, there's been a couple things that happened in the last couple days, which I think are incredibly important."
"There have been people that decompiled the entire, the entire big five. What is the big five? Adobe has five very, very important products in their product suite. Freeberg may know them all, but I just know Photoshop. There's like a bunch of others. Okay."
"And these folks decompiled it, distilled it, and published open source versions of these things. And Rust, yeah. And I was so shocked, because I was like, hold on a second, these are like profound products that have existed and have had moats for years, and now they're just open source versions."
"And then Nikita was, I think, trolling, and he was building on that idea, and just talking about GTA, because it's about to get released. I think GTA 6 is about to come out, and we'll reconstruct this."
"Why is this important?"
"I think that IP and software was effectively rendered worthless two days ago."
"Because if you think about it, when everything is headless, so when Elon makes his services headless, Benioff made Salesforce headless a few months earlier, when all of these things are headless, and you can wrap them in agent tools, and you can connect them via MCP to other experiences, the value of IP does no longer exist in software. It doesn't exist."
"And I think that probably is the most profound implication."
"So, like, even in our business, we would argue with some of our customers all the time, because we'll say, oh, we'll build this for you, we'll own the IP. They'll say, no, we want the IP. We don't know how to negotiate it. And two days ago, I walked into the 8090 guys, I said, who cares? Who cares? But before, it was always a negotiation and a set of trade-offs. Now it's a who-cares thing, where, like, the IP is irrelevant."
"It still takes a while for the machinery of a business, and frankly the market, to understand. But that is the most important thing that happened this week in AI, in my opinion."
David Sacks says Anthropic is run by a "religious cult" and is doing the very things it says it fears most, like opening a wet lab while warning of bio risk, and that there's an argument for sending in a federal SWAT team
"But I do think that there are potential externalities from having the number one frontier model company run by a religious cult."
"So, for example, they are opening a Claude-powered wet lab in the middle of San Francisco. But at the same time, they're warning of bio risk. At the same time, they say that we have to pace the frontier, they're extending the frontier. At the same time that they are saying that there's a risk of superintelligence growing beyond our control, they're programming it to grow beyond our control."
"I think there are potentially huge externalities to the way they approach these things. And it's, like, insane that they call these decisions "safety." These are not safe decisions. These are crazy decisions, it seems to me."
"This is why I've called this a schizophrenic company, is they appear to be not only hypocritical, but they're doing the things that they say they are most afraid of."
"Like, what is going on here? This is bizarre. And I don't know where this all leads. I do think there's an argument for a federal SWAT team to come blasting in there."
Readers added context they thought people might want to know
The open-source Adobe alternatives were clean-room reimplementations built from scratch, not decompiled Adobe code. The project states they were built from public specs and observed behavior with no proprietary code or assets. notebookcheck.net/Photoshop-Prem… explainx.ai/blog/photocraf…
Alexandre Kateb retweeted
An astrophysicist worked with Claude Science to create the first complete ultraviolet map of the sky. Astronomers have already created complete sky maps from radio to gamma rays. But large regions of the sky have never been observed in UV.
In today’s science blog, Brice Ménard explains how he guided Claude to find existing datasets, combine them, and fill in the gaps with statistical inference. The work would have taken humans weeks of work, but, with Claude, just took a few days while Ménard worked on other projects.
The resulting map is a useful teaching tool and an example of exactly the sort of slow, low-priority work which scientists never quite get around to, but which AI now makes feasible.
Read more: anthropic.com/research/the-m…
Alexandre Kateb retweeted
France: between the bond market and the barricades ft.trib.al/PhRAvqE
Alexandre Kateb retweeted
Le mouvement lycéen et étudiant signe le crépuscule du macronisme: les jeunes veulent se former et accéder aux études, et un pouvoir élitiste leur répond en réduisant de 20% le budget par étudiant entre 2017 et 2026 et en gérant la pénurie avec un algorithme opaque (Parcoursup)
Alexandre Kateb retweeted
CHART OF THE DAY: The cost of hiring a supertanker to ship oil from the Persian Gulf to the Far East has climbed further, reaching ~$1.3 million per day.
(It was less than $50,000 a day before the war started; and ~$1.1 million when I first published the chart in September)
Alexandre Kateb retweeted
China is getting closer to turning rocket launches into a process that begins on the ground before reaching the sky.
The “天弓” or “Heavenly Bow” project uses a superconducting electromagnetic track to accelerate the rocket to high speeds before its engines ignite, reducing the fuel needed during the most difficult stage of launch, while increasing payload capacity and lowering the cost of reaching orbit.
China has already conducted real tests of the technology in Sichuan and is working to adapt the Ceres-2 rocket to the system, with the goal of attempting the first electromagnetic-assisted orbital launch by 2028.