🚀 IA « anti-émotions » adoucit la voix des clients mécontents; voiture volante; crazy fish intelligence & more
Bonjour,
Vous recevez la 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 🔮.
Je m'appelle Thomas, co-fondateur de YeldaAI, qui développe des IA pour répondre au téléphone en langage humain pour les administrations et les entreprises. Plus d'infos sur moi en cliquant ici.
Et voici donc ma toute dernière sélection !
L’apéro
Les derniers épisodes du podcast Parlons Futur (Spotify, Apple, Youtube), notamment :
La superintelligence dans quelques milliers de jours dixit "Sam Altman" ? Est-ce-possible ?
Exemples choquants de comment l'IA pourrait révolutionner la médecine
Insolite : l'appli où vous êtes la star et parlez à vos millions de fans animés par l'IA
Le robot Optimus de Tesla est-il vraiment autonome ? Quels risques ?
Dernière vidéo bluffante de Boston Dynamics de leur robot humanoïde Atlas (vidéo)
Atlas is autonomously moving engine covers between supplier containers and a mobile sequencing dolly. The robot receives as input a list of bin locations to move parts between. Atlas uses a machine learning (ML) vision model to detect and localize the environment fixtures and individual bins
Unlike Tesla's Optimus bot, there are no prescribed or teleoperated movements; all motions are generated autonomously.
The robot is able to detect and react to changes in the environment (e.g., moving fixtures) and action failures (e.g., failure to insert the cover, tripping, environment collisions using a combination of vision, force, and proprioceptive sensors.
Cool video stopped at the right second: a robotic hand that can detach from its arm and then crawl around to grasp objects that would be otherwise out of reach
a nice reminder that robots don’t have to be constrained by our constraints
Amazing 40-sec video of a neural net in action showing how it recognizes what digit is being written by a user
Au Japon, SoftBank lance une IA « anti-émotions » qui adoucit la voix des clients mécontents (L'ADN)
Un « Emotion Canceling Voice Conversation Engine » pour protéger les employés des clients qui hurlent ou râlent sur eux
La technologie ne transforme pas les propos de l’appelant, elle modifie l’intonation pour la rendre moins agressive.
objectif de commercialiser sa nouvelle technologie d'ici 2026
Google CEO Says Over 25% of New Google Code Is Generated by AI (source)
One of the most dynamic flight we’ve seen from an eVTOL with a human on board (vidéo de 4 min)
eVTOL: electric Vertical Take Off & Landing vehicle
Big transitions are slow: The steam engine was invented in 1712... (source)
An observer at the time might have said: “The engine will power everything: factories, ships, carriages. Horses will become obsolete!”
But two hundred years later, we were still using horses to plow fields.
AI may not take 200 years to make itself felt, but it will still take some time before we know how to best harness it
After the invention of railroads, at first, the use of draft horses did not decline: it increased.
Railroads provide long-haul transportation, but not the last mile to farms and houses, so while they substitute for some usage of horses, they are complementary to much of it. An agricultural census from 1860 commented on the “extraordinary increase in the number of horses,” noting that paradoxically “railroads tend to increase their number and value.”
A similar story has been told about how computers, at first, increased the demand for paper.
British self-driving tech startup Wayve seems to be ahead of even Waymo and Cruise (MIT Tech Review)
Everything that Wayve’s cars do is learned rather than programmed. The company uses different technology from what’s in most other driverless cars. Instead of separate, specialized models trained to handle individual tasks like spotting obstacles or finding a route around them—models that must then be wired up to work together—Wayve uses an approach called end-to-end learning.
This means that Wayve’s cars are controlled by a single large model that learns all the individual tasks needed to drive at once, using camera footage, feedback from test drivers (many of whom are former driving instructors), and a lot of reruns in simulation.
Wayve has argued that this approach makes its driving models more general-purpose. The firm has shown that it can take a model trained on the streets of London and then use that same model to drive cars in multiple UK cities—something that others have struggled to do.
"We steer around roadworks, cyclists, and other vehicles stopped in the middle of the street. It starts to rain. At one point I think we’re on the wrong side of the road. But it’s a one-way street: The car has spotted a sign that I didn’t. We approach every intersection with what feels like deliberate confidence."
cars are just the start, says Rus. What Wayve is in fact building, he says, is an embodied model that could one day control many different types of machines, whether they have wheels, wings, or legs.
In America NASA, the US Space Force and other public institutions spent $73bn on space in 2023. In Europe the figure was less than $12bn. (The Economist)
ChatGPT Voice : 5 cas d'usage bluffants de l'IA vocale d'OpenAI, avec prompt et exemple audio d'1 minute (JDN)
1. Apprendre une nouvelle langue
2. Simuler un entretien d'embauche
3. Se faire assister en cuisine
4. Un conteur d'histoire pour enfant
5. Un guide touristique
La Ville d'Arcachon offre l'accès gratuit à ChatGPT à tous les habitants (source)
AI can help Polarized Groups of People Find Common Ground (MIT Tech Review)
an AI system using large language models could act as mediator in group discussions and help find points of agreement on contentious issues.
UK carbon emissions in 2023 were at their lowest level since 1879 (solar power expert Jenny Chase from BloombergNEF)
Can AI fix everything? wonders Steven Levy from Wired
"We don’t need a large language model to tell us war is hell and we shouldn’t kill each other. Yet wars keep happening."
" if a model from OpenAI, Google, or Anthropic came up with an injectable cure for Covid tomorrow, you know exactly what would happen. Large segments of the population will warn that it’s some insidious plot to wipe out everyone"
"Likewise, we already know how to mitigate the climate crisis, but we’re consuming more energy than ever."
"The human problem that AI will never solve is humanity itself, in all its glory and shame."
MIT spinoff Liquid debuts AI models not based on usual transformer architecture and are already state-of-the-art (source)
Liquid Neural Networks (LNNs) seel to make the artificial “neurons” or nodes for transforming data, more efficient and adaptable.
Unlike traditional deep learning models, which require thousands of neurons to perform complex tasks, LNNs demonstrated that fewer neurons—combined with innovative mathematical formulations—could achieve the same results.
Two Students Created Face Recognition Glasses. It Wasn’t Hard. (NYT)
“All the tools were there,” Mr. Nguyen said. “We just had the idea to combine them together.”
Meta glasses + Face detection software, which captures faces that appear on the livestream + A face search engine called PimEyes, which finds sites on the internet where a person’s face appears.
Chinese text-to-image model Kling ‘reunites’ users with their deceased relatives through animated photography (source and example videos)
Lol: Claude AI Gets Bored During Coding Demonstration, Starts Perusing Photos of National Parks Instead! (source)
A Tiny New Open-Source AI Model Performs as Well as Powerful Big Ones (MIT Tech Review)
[The Allen Institute for Artificial Intelligence (Ai2)] claims that its biggest Molmo model, which has 72 billion parameters, outperforms OpenAI’s GPT-4o, which is estimated to have over a trillion parameters, in tests that measure things like understanding images, charts, and documents.
large multimodal language models are trained on vast data sets containing billions of images and text samples that have been hoovered from the internet, and they can include several trillion parameters. This process introduces a lot of noise to the training data and, with it, hallucinations
In contrast, Ai2’s Molmo models have been trained on a significantly smaller and more curated data set containing only 600,000 images, and they have between 1 billion and 72 billion parameters. This focus on high-quality data, versus indiscriminately scraped data, has led to good performance with far fewer resources
Ai2 achieved this by getting human annotators to describe the images in the model’s training data set in excruciating detail over multiple pages of text. They asked the annotators to talk about what they saw instead of typing it. Then they used AI techniques to convert their speech into data, which made the training process much quicker while reducing the computing power required.
Another funny video: What happens when the Waymo driverless cars fail. 19 cars lined up and 4 minutes later
Here's What Happens When You Give People Free Money (Wired)
OpenResearch found that when it gave some of the poorest Americans $1,000 a month for three years with no strings attached, they put much of the money toward basic needs such as food, housing, and transportation.
But what amounted to $36,000 wasn’t enough to significantly improve their physical well-being or long-term financial health, researchers concluded.
“Cash offers flexibility and may increase agency to make employment decisions that align with recipients’ individual circumstances, goals, and values,” the researchers wrote. They may be “taking more time to find a job, taking a lower paying position that they find more meaningful, or simply taking a break.”
OpenResearch found “the total amount of work withdrawn from the market” was “fairly substantial” in its experiment.
researchers found “no effect” from the cash on several measures of physical health
A fun (and probably apocryphal) story about Milton Friedman once visiting China
He was shocked to see that, instead of modern tractors and earth movers, thousands of workers were toiling away building a canal with shovels. He asked his host, a government bureaucrat, why more machines weren’t being used.
The bureaucrat replied, “You don’t understand. This is a jobs program.”
To which Milton responded, “Oh, I thought you were trying to build a canal. If it’s jobs you want, you should give these workers spoons, not shovels!”
Animal intelligence: the most insane things fish do (source)
"You have a tank split in two parts:
if the fish gets in the compartment with a red circle, it will receive food, and food will be delivered in the other tank as well.
If the fish takes the blue triangle, this fish will receive food, but nothing will be delivered in the other tank.
So we have a prosocial choice and antisocial choice.
When there is no one in the other part of the tank, the male is choosing randomly.
If there is a male, a possible rival: antisocial — almost 100% of the time.
Now, if there is his wife — his female, this is a prosocial choice all the time.
"And now a question: Is it just because this is a female or is it just for their female?
Well, when they're bringing a new female, it’s the antisocial choice all the time.
Now, if there is not the female of the male, it will depend on how long he's been separated from his female.
At first it will be antisocial, and after a while he will start to switch to prosocial choices.
In some areas of military strength, China has surpassed America (The Economist)
China is also thought to churn out stealth fighters faster than America does.
The most striking progress by China has come in the area of hypersonic missiles, which fly and manoeuvre at more than five times the speed of sound.
thanks to big investments and extensive testing, China now has the world’s leading hypersonic arsenal. America is testing faster and more accurate missiles, says Mr Mulvaney. But China has already deployed multiple hypersonic-weapon systems.
Glydways, a new form of public transit (promoted by famous investor Vinod Khosla)
Far superior to self-driving vehicles as it increases street capacity 10x.
It's a bicycle-lane width self-driving system that doesn't ever get into a traffic jam, never stops at traffic lights, is point to point (no stops), personal transit.
It upends every feature of public transit!
Noah Smith on this once timeless situations that tech progress has solved (source)
“getting bored"
Like me, you're old enough to remember a thing called “getting bored.” Do you remember that? You’d just sit around and you're like, “Man, I’ve got nothing to do. I’m bored.” That emotion just doesn't exist anymore — I mean, very fleetingly for some people, but we've banished boredom from the world.
“getting lost"
Remember “getting lost?” If you walk into that forest, you might get lost? That doesn't happen unless you want to get lost, unless you don't take your phone. But the idea that, “Oh my God, I'm lost! I'm lost!” No, just look at Google Maps and navigate your way back.
Being lost and being bored are fundamental human experiences that have been with us for literally millions of years, and now they're just gone in a few years, just gone!
Remember when you didn't know what other places looked like?
You would think, “Oh, the Matterhorn, that’s some mountain in Switzerland, I can only imagine what that looks like.” And then maybe you'd look it up in an encyclopedia and see a picture of it or something. Now you just type it into Google Images, or Street View, or look at YouTube, look at a walking tour or something.
Remember not knowing how to fix things?
You just had no idea how to fix it. You could try to make it up, but really what you'd do is you'd call someone who was handy with stuff who had this arcane knowledge, and this wizard would fix your cabinet, or your dresser, or whatever, your stereo. Now we have all these Youtube tutorials...
Partagez cette newsletter par Whatsapp en cliquant ici ❤️
Si vous appréciez cette première moitité de l’apéro gratuite, n’hésitez pas à prendre 3 secondes svp pour transférer cet email ou l’envoyer à ne serait-ce qu’un contact par Whatsapp en cliquant ici 🙂
Et si on vous a partagé cet email, vous pouvez cliquer ici pour vous inscrire et ne pas manquer les prochains
Je vous invite à vous abonner à la version payante pour recevoir lundi les résumés de ces longs articles le plus souvent issus des plus grands journaux (et accéder aux archives) en cliquant sur le bouton “Upgrade to paid/Subscribe now” plus bas :
Do we have enough minerals for the energy transition?
How much of a logistical minefield would invading Taiwan be for China?
How China’s EV Boom Caught Western Car Companies Asleep at the Wheel
Les dernières newsletters :
L’addition ?
Cette newsletter est gratuite, si vous souhaitez m'encourager à continuer ce modeste travail de curation et de synthèse, vous pouvez prendre quelques secondes pour :
transférer cet email à un(e) ami(e) ou partager par whatsapp
et ici pour la version 100% en anglais
étoiler cet email dans votre boîte mail
cliquer sur le coeur en bas d’email
Un grand merci d'avance ! 🙏
Ici pour s’inscrire et recevoir les prochains emails si on vous a transféré celui-ci.
Merci, et à bientôt !
Thomas