DataThoughts (@wishful_data) profile picture
Joined March 2023
Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)Photo by DataThoughts (@wishful_data)
I dislike ground floor primary bedrooms, just throwing that out there in case an architect sees this
1
66
Is there a @FrameworkPuter for TV’s? I hate “smart TVs” especially @Samsung
67
DataThoughts retweeted
Elon Musk avait dit un truc qui m'avait marqué sur l'allocation de ressources. En substance : passé un certain niveau de richesse, l'argent n'est plus de la consommation, c'est de l'allocation de capital. Cette phrase change tout. L'économie, dans le fond, c'est juste un problème d'allocation. Tu as des ressources finies et des usages infinis. Qui décide où va quoi ? Imagine une cour de récré. 100 enfants, des paquets de cartes Pokémon distribués au hasard. Tu laisses faire. Très vite, un ordre émerge. Les bons joueurs accumulent les cartes rares, les collectionneurs trient, les négociateurs trouvent des deals. Personne n'a planifié. Et pourtant chaque carte finit dans les mains de celui qui en tire le plus de valeur. Le système maximise le bonheur total de la cour. C'est ça, la main invisible. Maintenant fais entrer la maîtresse. Elle trouve ça injuste. Léo a 50 cartes, Tom en a 3. Elle confisque, redistribue, impose l'égalité. Trois effets immédiats. Les bons joueurs arrêtent de jouer, à quoi bon. Les mauvais n'ont plus de raison de progresser, ils auront leur part. Les échanges s'effondrent. La cour est égale, et morte. Elle a maximisé l'égalité, elle a détruit le bonheur. Le problème de la maîtresse, c'est qu'elle ne peut pas avoir l'information que la cour avait collectivement. C'est le problème du calcul économique de Mises, formulé en 1920. L'URSS a essayé de le résoudre pendant 70 ans avec le Gosplan. Résultat : pénuries, queues, effondrement. Pas parce que les Soviétiques étaient bêtes, parce que le problème est mathématiquement insoluble en mode centralisé. Quand Musk a 200 milliards, il ne les consomme pas, il les alloue. SpaceX, Starlink, Neuralink, xAI. Chaque dollar est un pari sur le futur. Et lui a un track record. PayPal, Tesla, SpaceX. Il a démontré qu'il sait identifier des problèmes immenses et y allouer des ressources avec un rendement spectaculaire. L'État aussi a un track record. Hôpitaux qui s'effondrent, éducation qui décline, dette qui explose, services publics qui se dégradent malgré des budgets en hausse constante. Le marché identifie les bons allocateurs, la politique identifie les bons communicants. Le profit n'est pas une finalité, c'est un signal. Il dit : tu as alloué des ressources rares vers un usage que les gens valorisent suffisamment pour payer. Plus le profit est gros, plus la création de valeur est grande. Quand Starlink est rentable, ça veut dire que des millions de gens dans des zones rurales ont enfin internet. Quand un ministère est en déficit, ça veut dire qu'il consomme plus qu'il ne produit. L'un crée, l'autre détruit, et on appelle ça redistribution. Dans nos sociétés il y a deux catégories d'acteurs. Les entrepreneurs et les bureaucrates. L'entrepreneur prend un risque personnel pour identifier un problème, mobiliser des ressources, créer une solution. S'il se trompe il perd. S'il a raison, ses clients gagnent, ses employés gagnent, ses fournisseurs gagnent, l'État collecte des impôts. Il est la cellule de base du progrès humain. Le bureaucrate ne prend aucun risque personnel. Son salaire est garanti. Au mieux il maintient une rente existante. Au pire il la détruit par excès de réglementation, mauvaise allocation forcée, incitations perverses qui découragent ceux qui produisent. Mais dans aucun cas il ne crée. Regarde les 50 dernières années. iPhone, internet civil, SpaceX, Tesla, Google, Amazon, Stripe, mRNA, ChatGPT. Toutes des inventions privées, portées par des entrepreneurs, financées par du capital risque. Pas un seul ministère n'a inventé quoi que ce soit qui ait changé ta vie au quotidien. La France est devenue le laboratoire mondial de la dérive bureaucratique. 57% du PIB en dépenses publiques, record absolu. Une administration tentaculaire, une fiscalité qui pénalise la création de richesse. Résultat : décrochage face aux États-Unis, à l'Allemagne, à la Suisse. Fuite des cerveaux. Désindustrialisation. Dette qui explose. Et le pire c'est que la mauvaise allocation s'auto-renforce. Plus l'État prélève, moins les entrepreneurs créent. Moins ils créent, moins il y a de base fiscale. Plus l'État s'endette et taxe. Boucle de rétroaction négative parfaite. La maîtresse pense qu'elle aide, et chaque année la cour produit moins. Dans nos sociétés, ce sont les entrepreneurs, toujours, qui font avancer la civilisation. Les bureaucrates au mieux maintiennent une rente, au pire la détruisent. Aucune société n'a jamais progressé en taxant ses créateurs pour subventionner ses gestionnaires. La question n'est jamais qui a combien. C'est qui alloue le mieux la prochaine unité de ressource pour maximiser le futur de l'humanité. La réponse depuis 200 ans n'a jamais changé. Ce ne sont pas les fonctionnaires.
3,217
16,117
1,898
60,809
81,825,410
DataThoughts retweeted
For the record. President Trump’s self‑styled “Republican New Deal” marks a rupture with the mid‑century Keynesian consensus that has dominated central banking for nearly a century. It challenges the idea that technocrats at the Federal Reserve can fine‑tune demand and rescue the economy every time asset prices wobble, and instead re‑centers production, work, and national capacity as the core of economic strategy. Kevin Warsh’s charge that the Fed “should abandon the dogma that inflation is caused when the economy grows too much and workers get paid too much” goes straight at the priesthood of Phillips‑curve Keynesianism: the reflexive view that tight labor markets are a pathology, and that the real economy is something to be cooled rather than cultivated. Anyone who has taken a serious history of economic thought course knows that Keynes wrote The General Theory for a very specific use case: the catastrophic unemployment and demand collapse of the 1930s, when markets plainly failed to self‑correct. He did not write it as a universal operating system for every macroeconomic context from stagflation in the 1970s to supply shocks, to C-19, and geopolitical competition in 2026. Even Keynes himself had misgivings about the “general” label, while his publisher had every incentive to oversell the book as a sweeping theory to move copies. The tragedy is that successive generations of central bankers and academic Keynesians treated that contingent Depression‑era playbook as a timeless catechism, long after its empirical foundations had crumbled. Warsh’s deeper critique is that the Fed has become a de facto fiscal and industrial policymaker, distorting price signals, inflating asset values, and privileging Wall Street over Main Street through repeated, discretionary interventions. The “committee to save the world” era of Bernanke, Yellen, and Powell elevated the central bank into an unelected macroeconomic government, blurring every line between monetary, fiscal, and industrial policy while pretending this was neutral technocracy. That model is now colliding with a more Hamiltonian view: money and credit should serve productive, strategic investment, not perpetual asset‑price protection and demand management. This is where the shift from Keynesian to supply‑side and Hamiltonian thought matters for central banking in 2026. A genuinely modern framework cannot be built on the assumption that growth and wage gains are inherently inflationary, or that every downturn demands another round of balance‑sheet experimentation. It must accept that in an era of supply constraints, geopolitical fragmentation, and national industrial strategies, the central bank’s job is narrower but more demanding: sound money, stable credit, and respect for the distinction between monetary and fiscal power. Diversity of economic thought at the Fed is not some ideological heresy; it is the minimum requirement for intellectual honesty after the serial failures of one school of thought. Keynesians will bristle at this, because it forces an uncomfortable admission: The General Theory is not “general” in 2026. It is a brilliant Depression‑era intervention that has been stretched far beyond its domain of validity, weaponized to justify an ever‑expanding central bank that socializes downside risk while ossifying the real economy. The emerging paradigm says: enough. The age of single‑school macroeconomics sitting atop the Fed is over. Diversity of frameworks, Hamiltonian, supply‑side, and market‑oriented, will not weaken the institution; it is the only way to restore its legitimacy and reconnect monetary policy to the real, productive economy it is supposed to serve.
86
257
38
843
49,382
Clawdbot is a great reminder of how early we are, it’s literally just an agent and people think that it’s novel
2
122
Ok Kiro is sweet, nice work @awscloud - love the spec driven approach, seems to really keep the ai on rails
2
234
Requiring updates to play a game that can be played offline is ridiculous @Xbox
2
164
Removing Rayon made my code 10% faster
194
Python is like candy - it tastes good, is easy to consume, and there are lots of options to choose from, but all that sugar comes at a cost. Rust is like a healthy meal - harder to get, keeps you lean, but won’t make you feel like shit.
2
153
My main benchmark for SOTA LLMs has been how well it can translate novel ML research papers into efficient Rust implementations.
124
Paginated deduplication with LSH for Minhash signatures works pretty well for embedding preprocessing. Couple that with some post-embedding HNSW groupings and your LLM context gets pretty tidy
101
I think LLM benchmarks are bullshit. This is like S&P ratings during the ‘08 financial crisis. UI slop can be cool but we should be measuring how they solve HARD PROBLEMS, not integrations.
2
2
139
Ok I called this months ago: teapaw.org/kimmonismus/status/192…
Had this idea, and @grok fleshed it out: AI product placement in shows and movies involves the use of artificial intelligence to dynamically insert or modify brand mentions and product placements. Here's how it typically works: - **Dynamic Insertion:** AI algorithms analyze dialogue or visual elements where generic products or brands are mentioned or shown. For example, if a character says, "I need to buy a new truck," the AI could change it to, "I need to buy a new Ford," based on sponsorship deals or advertising agreements. - **Real-Time Customization:** In streaming platforms, this technology can customize product placements for different viewers or regions. This means viewers in one country might see a local brand, while viewers in another see an international one, enhancing market relevance and potentially increasing advertising revenue. - **Ethical and Creative Considerations:** - **Authenticity:** There's a challenge in maintaining the authenticity of the narrative. Overuse or poorly integrated AI product placements can feel unnatural or disrupt the viewer experience. - **Consent and Rights:** Actors, writers, and directors might have rights or contractual stipulations regarding changes to their work, which could complicate the application of AI in this way. - **Viewer Perception:** There's a risk of audience backlash if product placements are seen as too invasive or manipulative. Transparency about AI modifications might be necessary to maintain trust. - **Technical Implementation:** - **Speech Synthesis:** For dialogue, AI would need sophisticated speech synthesis to match the actor's voice, tone, and accent. - **Visual Editing:** For visual placements, AI would require capabilities in image recognition and manipulation to seamlessly integrate products into scenes. - **Legal and Regulatory Issues:** Advertising laws vary by country, and using AI for product placement might run afoul of regulations concerning transparency in advertising, especially if not disclosed to viewers. - **Future Potential:** As AI technology advances, we might see more nuanced applications where product placements are not only brand-specific but also tailored to the viewer's interests, potentially leading to a new form of interactive or personalized entertainment. The integration of AI in product placement could revolutionize how media monetizes content, but it requires careful balancing to ensure it does not detract from the storytelling or viewer experience.
1
1
167
Ok just did a quick codex trial run by having it try to generate some unit tests for my rust mcp sever. It knocked out ~10 basic unit tests across 3 files. I then ran `cargo test` and it created 17 build errors. I said please verify your tests and fix the build errors. It thought for 300 seconds before I terminated it. Maybe I’ll try it again in a few months.
1
93
Seeing overly complicated design patterns become anti patterns is a beautiful thing
46
The rise of AI has made people’s lack of agency extremely apparent - I see people on here asking for a piece of software that does this or that, and its like you can just build it yourself…
154
Folded and tried FastMCP - it works great, and now so does my MCP
1
156
Replying to @Starlink
@Starlink I would love to be able to toggle my priority data off during non essential times or for non essential devices on my network
1
162
Like I want to be able to download the 100GB @assassinscreed game without using the priority data I use for work
142
My first computer was a Compaq Presario running Windows 95. Rocked a Pentium 75 MHz CPU, 8 MB RAM, and basic 2 MB graphics—enough to play Minesweeper like a champ!
1
106
Pretty wild that I have access to AGI (Grok 3), on an airplane. Just hurdling through the air, chatting with a thing more intelligent than most of the other passengers. @xai @elonmusk
1
84