Parlons Futur

L'IA peut-elle créer de faux souvenirs ? Retours et échanges suite à une nouvelle étude

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/

 

Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

L'ultra-personnalisation de l'expérience client/citoyen permise par la data et l'IA est-elle souhaitable ?

• Pierre Bellanger, fondateur et CEO du groupe Skyrock vient de publier un nouvel essai « La Granularité »
  • Il a développé ce concept lors d’une conférence intitulée Un futur singulier au ministère des Armées en 2021.
  • "Aujourd’hui, la société est gouvernée par la statistique, mais demain, elle sera régie par les données."
  • "Lorsqu’on dispose de peu d’informations sur chaque individu, on administre par catégories. En revanche, lorsqu’on possède une grande quantité d’informations sur chacun, on gère individu par individu : c’est cela, la granularité."
  • En voici un résumé de 12 min au format conversation podcast généré via l'IA avec l'outil de Google NotebookLM, bluffant !
  • Et sinon, la conversation exclusive entre vrais humains, lui et moi, à retrouver sur mon podcast Parlons Futur sur Apple Podcast et Spotify)

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/ Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

Que penser du robot Optimus de Tesla ? Quels risques dans un monde de robots autonomes et intelligents ?

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/ Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

Insolite : l'appli où vous êtes la star et parlez à vos millions de fans animés par l'IA

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/

 

Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

Exemples choquants de comment l'IA pourrait révolutionner la médecine

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/

Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

Pourquoi OpenAI bloque-t-il le Voice Mode de ChatGPT en Europe ?

Abonnez-vous à la newsletter Parlons Futur ici : https://www.parlonsfutur.com/

 

Une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech 🤖, science 🔬, éco 💰, géopolitique 🌏 et défense ⚔️ pour mieux appréhender le futur 🔮.

La superintelligence dans quelques milliers de jours dixit "Sam Altman" ? Est-ce-possible ?

Sam Altman a dit : "It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there."

source: https://ia.samaltman.com/

 

Alors possible ? Comment ? Quelles contraintes ? Quels risques ?

 

OpenAI Scale Ranks Progress Toward ‘Human-Level’:

• Problem Solving Level 1 (where it think it's at): the kind of AI available today that can interact in conversational language with people
• Level 2 (where we're entering): “Reasoners.” This refers to systems that can do basic problem-solving tasks as well as a human with a doctorate-level education who doesn’t have access to any tools.
• Level 3: Agents,” referring to AI systems that can spend several days taking actions on a user’s behalf.
• Level 4: AI that can come up with new innovations.
• Level 5: AI that can do the work of an organization (same as AGI or superintelligence)

Agathe & Thomas : GPT-4o ; la fin des devoirs ? ; confier son Tinder à l'IA ?

Nouveau format dans lequel Agathe interroge Thomas, cofondateur de Yelda, sur certaines news qu'il a partagées dans le dernier numéro de la newsletter Parlons Futur 🚀"The best computer interface I’ve ever used" ; Musk Vs LeCun Twitter fight ; IA pour draguer, faire la guerre & more

Au menu :

1. GPT-4o : qu'en retenir
2. 00:11:20 : GPT-4 peut faire quasiment tous les devoirs de jeunes : qu'est-ce qu'on dit aux parents, aux professeurs ?
3. 00:25:20 : Expérience Tinder : une IA a discuté avec une trentaine de femmes pendant 24h, elles n'ont pas vu la différence.

Vous abonner à Parlons Futur : https://www.parlonsfutur.com/

Agathe Rocher : https://www.linkedin.com/in/agatherocher/

Thomas Jestin : https://www.linkedin.com/in/thomasjestin/

Yelda, l'IA vocale de la relation client : https://www.yelda.fr/

 

Résumé du livre The Coming Wave, écrit par un co-fondateur de DeepMind

Conversation with Casey Handmer: how Terraform Industries will make the world carbon neutral

Two years ago, Casey Handmer quit working on GPS science at NASA JPL to start Terraform Industries, a wild idea backed by visionary investors (including Stripe co-founder and CEO) to bolt our growing solar abundance to growing hydrocarbon scarcity.

His plan? To make the world's cheapest green hydrogen, the world's cheapest direct air capture CO2, and cheap natural gas.

This is my conversation with Casey! It's quite long, but I've detailed the timestamps below, Casey elaborated on many important issues, make sure to give them a listen :)

Recorded on Oct 12th.

And mind the two questions about Mars at the end :)

• 00:00:28: Casey Handmer's background in science and industry and what led him to found Terraform Industries

• 00:02:30: Why Hyperloop One, which Casey worked on, didn't work as expected

• 00:04:38: Why "Terraform Industries" as the company name?

• 00:09:45: Terraform Industries' philosophy

• 00:13:20: How Terraform Industries is a fundamental bet on cheap solar PV electricity and why it will get ever cheaper for a long time

• 00:18:08: What the cost of solar PV electricity needs to be for Terraform Industries' green e-methane to start being competitive with fossil methane

• 00:20:20: What it needs to be for green e-kerosene to be competitive with fossil kerosene, in order to make aviation carbon-neutral

• 00:21:30: Casey explains how we may even be able to make food thanks to cheap solar, air and water

• 00:24:00: Can total cumulative solar PV deployment continue to double at current pace (approx. once every 2 years) long enough to have enough capacity to displace all fossil fuels?

• 00:30:15: Why has Casey been sharing his plans very openly so far? 

• 00:32:50: Casey shares an interesting analogy with the electric car industry

• 00:36:00: Can Europe compete with China on the EV market? do tariffs help?

• 00:37:55: How does Terraform industries work? What are the three subsystems of its first product?

• 00:41:30: Why does Terraform Industries want low-efficiency systems?

• 00:46:50: What should the be the cost of solar PV electricity for Terraform Industries' green e-methane to start being competitive?

• 00:49:00: What could be the market size for such machines? And as they're low tech and will be copied, what minimum market share could Terraform Industries hope to retain ultimately?

• 00:50:30: Does Terraform Industries intend more on selling the machines or operating the machines and selling green e-hydrocarbons?

• 00:53:25: Is what Terraform Industries is doing easier than coming up with and industrializing the Haber-Bosch process? (that helps fix nitrogen as ammonia from the air and make fertilizer, letting us feed billions more than if we didn't have it otherwise)

• 00:58:35: Hear Casey say which percentage of solar arrays, easily, will be used to make synthetic fuel 10 years from now, shocking!

• 00:58:52: What's the status right now for Terraform Industries? What are the next milestones?

• 01:01:28: What risks is Terraform Industries facing?

• 01:04:50: What about methane leaks? Especially as Casey Handmer expects world methane consumption to go up as methane gets cheaper and carbon-neutral

• 01:07:20: As Terraform Industries actually makes green hydrogen (before reacting it with captured CO2 to make methane CH4), why not stop there and provide hydrogen to the "hydrogen economy" that people see about to boom? For industrial heat, aviation, shipping etc.

• 01:13:20: Why is Casey Handmer fascinated with airships?

• 01:19:38: What about nuclear fusion? Could it help do without solar plants? (Helion talks about 1MW fusion plant the size of a container as opposed to 5-10 acres for a 1MW solar PV plant)

• 01:28:00: Why does Casey Handmer think that ultimately most of solar PV electricity will be used to synthesize green hydrocarbons for aviation (which will hence boom) as opposed to every other use combined.

• 01:37:00: Can Starhip be a thing when it comes to point-to-point transport on Earth?

• 01:39:55: Why taxing carbon doesn't work

• 01:44:00: Why current carbon capture and storage approaches can't scale to any meaningful scale

• 01:46:20: How does AI progress impact what Terraform Industries is doing? What would AGI do?

• 01:49:05: Will what Terraform Industries is doing be commoditized? Making it hard to find investors? 

• 01:52:35: Stratospheric Aerosol Injection : one of the two other important things we need to be doing to fix the climate crisis (beyond making all we do carbon-neutral for the atmosphere)

• 01:57:10: The other thing, "enhanced weathering", the best way to remove legacy CO2 from the atmosphere (possibly at $20/ton of CO2)

• 02:00:10: Casey Handmer's spiciest takes on energy (why batteries will ultimately bankrupt most long power lines, and also why it won't make sense to build more fission reactors)

• 02:08:48: What about the interseasonal intermittency problem with solar PV?

• 02:11:23: Hear Casey say why in the long term even methane pipelines will struggle to operate profitably as methane will be synthesized locally, even in temperate latitudes

• 02:19:30: Sending humans to Mars means bringing with us our microbes. As Mars could be habitable and inhabited by local microbes, even near the surface, we could ending confusing our search for life there, as Earth and potential Mars life may be related. Isn't that a reason to hold on for now and just send our robots which present much less of a risk?

• 02:24:24: We make all these plans to settle Mars, but what if Mars gravity is a non-starter for humans? Shouldn't the first thing be for us to build ae a space station that can be spun to recreate Mars gravity so as to test its effect on us, before anything else?

• 02:28:50: What surprised Casey the most moving from Australia to the US. About California and Texas.

• 02:32:00: Books recommended by Casey (see below for references)

• 02:32:45: About Elon Musk

 

Casey Handmer's blog articles we discussed (highly recommended):

• How to Produce Green Hydrogen for $1/kg

• Radical Energy Abundance

• Grid Storage: Batteries Will Win

• The Terraformer Mark One

• We should not let the Earth overheat!

• Terraform Industries Whitepaper 2.0 (Jan 2023)

• Is nuclear power a solution to climate change?

 

My explainer: Green, On-Demand, Abundant Energy by 2050 is Definitely Possible: Here's the Plan with the direct link to section 2 focussing on what Terraform Industries is doing

 

Discussed in the conversation:

• The Leuna factory in Germany, built by BASF in 1916, that Casey visited and mentioned at 00:54:40

• The Zeppelin Hidenburg scene in Indiana Jones and the last crusade (1989)

• See on Google Earth an open quarry by the sea, south east of Santa Catalina Island, opposite Los Angeles, mentioned when discussing enhanced weathering as the best solution to scrub the atmosphere of legacy CO2.

• Recommanded Books: 

  • Pieces of the Action, by Vannevar Bush (1970)

  • The Alchemy of Air: A Jewish Genius, a Doomed Tycoon, and the Scientific Discovery That Fed the World but Fueled the Rise of Hitler, by Thomas Hager (2009)

 

Useful resources

• Casey Handmer's personal blog site

• Terraform Industries' website

• Terraform Industries' blog site

• Find Casey Handmer on Twitter

 

À propos de l'éditeur du podcast Thomas Jestin :

• Ma newsletter Parlons Futur : une fois par semaine au plus, une sélection de news, mêlant sources anglophones et francophones, résumées en bullet points sur des sujets tech, science, éco, géopolitique et défense  pour mieux appréhender le futur.
• La liste de mes plus de 50 essais à ce jour dont certains publiés dans le Journal du Net, Les Echos (ici), l'Express.
• Me retrouver sur Twitter : twitter.com/thomasjestin
• Plus d'infos sur moi
• Me contacter : parlonsfutur@substack.com