←
Fonss1,856 posts
Fonss avatar

Fonss

@fonss87

Ingeniero Informático. Te monto Tu Equipo IA. Descuento en Tesla: Supercarga gratis 1.000 kms con mi enlace de recomendación. ts.la/alfonso491177

◷ Joined July 2016▧ 46 Photos and videos
Fonss avatar
Fonss@fonss87·20h

Tesla is putting a €3B Cybercab factory at Giga Berlin: construction from H1 2027, production in 2029, sized for up to 1M Cybercabs a year, per Berlin plant chief André Thierig. That's capacity built before Europe has approved driverless cars at scale. A bet on regulators as much as on the robotaxi. Source: @AndrThie x.com/AndrThie/status/212236…

0:36

If you ever wondered, what we would be doing at Giga Berlin with all the land that we have… We are expanding! Cybercab production is coming to Giga Berlin! Today we announced a first investment of around € 3b in a new manufacturing building with an anticpated start of production in 2029! It will have the most advanced technology of any Tesla factory! And we are hiring 👉 tesla.com/careers/search/?si…

Post attachment
Fonss avatar
Fonss@fonss87·Nov 17

The real story in NVIDIA's GR00T 2 isn't the robot, it's the data: 100,000 hours of first-person human video. NVIDIA reports 71% average success across 50 tasks with one demo per task, 95% with five, and skills carrying over to different robot bodies. Foxconn is already applying it to server manufacturing. Footage: NVIDIA

0:35
Fonss avatar
Fonss@fonss87·Nov 16

Tesla's week in Physical AI: Cybercab opens to select buyers ($29,990 + $299/mo), Robotaxi hits San Antonio and goes 24/7, Finland approves FSD (Supervised), 50 Semis head to Chicago, Cybertruck extended to end of summer 2027. Sources on screen.

0:50
Fonss avatar
Fonss@fonss87·Nov 13

The glove matters more than the camera. In most grasps your fingers are hidden behind the object, so video alone guesses the hand pose. But a human hand isn't a robot hand. Does glove data transfer as is, or does every robot need its own retargeting?

Capturing how humans actually move is a critical aspect of humanoid training. That's what @GenrobotAI showed at CoRL 2026 with its new data acquisition system: a first-person capture rig and a dexterous data glove that record how we see, move, and handle objects, all at once. Vision, motion, and touch in one stream. Human data is the foundation for Physical AI.

0:12
Fonss avatar
Fonss@fonss87·Nov 11

10 million hours sounds huge until you split it across tasks. A few hundred skills, each needing varied homes, lighting and objects, and coverage per task gets thin fast. Will robot data scale by hours, or by how many different kitchens it comes from?

Humanoid robots: I need more, much more data… I just watched an interview with the Memories AI team. One thing really stood out to me: by next year, Physical AI data collection could be entering the 10-million-hour range. The entire field is desperately hungry for real-world data. Just a few months ago, the egocentric data volumes being discussed across the industry were still around 1 million hours. More recently, we've been seeing figures to 2 million hours. But that's still not enough. Even 10 million hours may not be enough... Across the Pacific, at the World Internet Conference in Wuzhen, China, Li Shuai, deputy director of the AI Center at China's Ministry of Industry and Information Technology's Fifth Electronics Research Institute, put the gap into perspective: The core supply-demand gap in embodied AI data is roughly two orders of magnitude. Effective data stock is at the million-hour level, while leading companies estimate their demand at 100 million hours. It's pretty clear that demand for data is far ahead of supply at the upstream end of Physical AI. This is becoming a genuine must-have business. Look at what's happening-- On one side, companies are building dedicated humanoid robot data-collection and training facilities. China alone has more than 100 publicly disclosed embodied AI data centers and training bases. and like the NEURA's RoboGym with TUM in Germany, Apptronik's Robot Park in the US, and Tesla's Optimus Academy are all part of this push… On the other side, NVIDIA, Google DeepMind and World Labs are pushing world models, synthetic data and simulation to expand the training pool. NVIDIA generates synthetic robot trajectories from human demonstrations. Google's Gemini Robotics 2.5 combines real-world demonstrations with simulated practice, while World Labs' Marble generates 3D environments for robot training and testing. Then there's another layer: companies like Memories AI are building platform-based, one-stop data infrastructure to collect, organize and supply first-person real-world data at scale. These three routes may look like different technical bets, but I think they will increasingly converge.We're already seeing this in models like Google's Gemini Robotics 2.5, NVIDIA's GR00T 2 and Skild AI's S1. They all bring together different combinations of human egocentric video, wearable demonstrations, robot teleoperation and simulation. The boundaries between these data pipelines are becoming less distinct. And there's another side to this race(this sounds rather unhumanistic)--access to real industrial environments, frontline workers, skilled technicians and service professionals around the world. Their practical experience is an enormous source of knowledge that robots need to learn from. I've said this more than once… This really is the gold rush for Physical AI data. But data is not the goal.

0:26
OpenAI avatar
⟳ Fonss retweeted
OpenAI@OpenAI·Nov 10

Now in beta: Businesses in ChatGPT. Claim your business, manage your information, and get insights on how people engage with you. Eligible businesses can also build a Business Agent to answer customers’ questions and help them explore what you offer. openai.com/index/introducing…

0:49
NVIDIA AI avatar
⟳ Fonss retweeted
NVIDIA AI@NVIDIAAI·Nov 6

Turn a single image into a 3D world you can explore with Lyra 2.0. By combining video generation with 3D reconstruction, our researchers create environments along a camera path you choose. Lyra draws on earlier views and learns to correct accumulating visual errors, helping the world stay consistent as you explore and return to familiar places. Paper, code, and demos: nv-tlabs.github.io/Project-L…

0:35
Fonss avatar
Fonss@fonss87·Nov 6

What's new here isn't that a robot reads the news. Chinese media already ran AI hosts on screens for years. The real product is the body: a persona that can leave the desk and walk a forum floor. Does physical presence change what audiences trust, or is trust still just the script?

Future still got here too fast. Xinhua News just hired two biomimetic humanoid news anchors from UBTECH’s UWORLD U1. They’ll host the reports. Chinese media has used AI digital hosts for years --daily news, short-video anchors. Now those digital people have bodies...Personality and memory can be loaded in, they can step off the screen and talk to people. Well, I suppose we’ll have to get used to humanoid robots serving as news anchors and even hosting forum events...

0:15
Md Ismail Šojal 🕷️ avatar
⟳ Fonss retweeted
Md Ismail Šojal 🕷️@0x0SojalSec·Nov 2

You can now Run Uncensored DeepSeek-V4.1-Flash- 1017 locally on a 12 GB GPU with zero refusals consumer hardware.

0:17

You can run locally Uncensored DeepSeek-V4.1-Flash-1017 on a 12GB GPU on consumer hardware. - Refusals collapsed from 85% to 0% on JailbreakBench (even during thinking). - Routed experts stayed byte-identical. - Reasoning held: GPQA unchanged, AIME slightly up. - Runs via Strata with 64 GB RAM offload SSD. - huggingface.co/AtomicChat/De…

Post attachment
Fonss avatar
Fonss@fonss87·Oct 30

Going from 1 AM to 2 AM isn't a driving problem, it's a depot problem. The dead hours are when a fleet gets charged, cleaned and checked, so each hour added pushes that work into rolling shifts while other cars keep driving. What breaks first on the road to 24/7: charging, cleaning or remote support?

Tesla has announced that their robotaxi service now runs until 2 AM (from 1 AM before). Getting close to 24-hour operation! Other updates: • Trip details can now be shared with your friends • Rear view camera feed now available during your ride • New Calm Mode: dims screen & sets cabin lighting to a "relaxing state." • Grok now in Model Y robotaxi Accessibility updates: • There is now haptic feedback to assist users in finding their Cybercab/Model Y robotaxi • Improved vehicle locating arrow in the app • Music Control via Screen Reader Been hard to keep track of all the robotaxi service updates lately lol. x.com/robotaxi/status/211602…

Post attachment
Agus | Experto en Youtube Automatizado avatar
⟳ Fonss retweeted
Agus | Experto en Youtube Automatizado@AgusYoutubeAuto·Oct 27

🚨NICHO REVELADO🚨 Un canal en ESPAÑOL con solo 14 vídeos. Empezó el 6 de octubre y NexLev ya le estima más de 5.000 $ al mes. 👇 En el comentario te digo por qué este tema engancha tanto. 📌 Se llama Ricos en Apuros: casas y edificios de lujo que se están viniendo abajo. Los datos (NexLev) 👇 ➡️ Primer vídeo: 6 de octubre ➡️ 14 vídeos · monetizado ➡️ Unas 2,1 millones de visitas al mes ➡️ NexLev estima unos 5.100 $ al mes ➡️ "Las villas de lujo de $50 millones de Sinaloa se derrumban en el Pacífico — y nadie puede detenerlo" → 696k ➡️ "Reñaca: edificios de $50 millones se deslizan hacia el mar — y Chile no los ayuda" → 511k Tres semanas. Monetizado. ¿Lo habías visto? 👇

Post attachment
VynqorxeAI avatar
⟳ Fonss retweeted
VynqorxeAI@Vynqorxe·Oct 28

Holy sh*t... 14 days ago this account didn't exist 🤯 Today: → 30M+ views → 232K followers → 10 posts → one hit: 15.1M → 41,509 votes cast by viewers No actor. No voice-over studio. No writers' room. Just an illustrated heist story: you stole $10M, and the comments decide what you do next. Every part teaches one real money rule: → $10M in hundreds = 100,000 bills → bait money and serial numbers → deposits over $10K get reported to the government → splitting them up has a name: structuring Viewers don't watch a lesson. They live it and vote. The creator makes money from the story itself: Founder's Pass ($15–20), business cameos ($150–$1,000), brand sponsorships, and a budgeting app at the end of the funnel. My estimate: $30–50K in 14 days. That's my guess, not his report. Made on Higgsfield. Now imagine the same format, but with your product at the end 👀 Follow me so you don't miss the next breakdown. The next one will be even crazier 🔥

1:27

Ernesto Lopez @ErnestoSOFTWARE Oct 21 Article How we automated a slideshow farm 10M+ views/mo ran by AI agents we generated 10M+ views on TikTok without filming a single video this is just one of our accounts we have more with a few million views that go way beyond 10M+ views no creators & no editors we

frontman🪖 avatar
⟳ Fonss retweeted
frontman🪖@frontm_an·Oct 28

a 16-year-old reportedly made $38k in a single month running faceless youtube channels with ai. no camera. no showing his face. no recording his own voice. here’s the system behind it: step 1: find a niche → use claude to identify video ideas people are already searching for. step 2: write the script → generate hooks, full scripts, and key moments that keep viewers watching. step 3: create the voiceover → turn the script into natural-sounding ai narration. step 4: produce the visuals → combine ai-generated visuals, b-roll, and simple animations into a complete video. step 5: batch and schedule → review the videos, prepare multiple uploads, and schedule a week of content in one sitting. the result? one person can manage multiple channels without handling every task manually. as the channels grow, youtube ad revenue and sponsorships can open up additional income streams. the bigger opportunity is using ai to simplify production and build a repeatable content system. but the tools alone won’t guarantee results. originality, video quality, audience retention, and niche selection still matter. the goal isn’t just to make videos with ai. it’s to build a system that consistently creates content people want to watch. bookmark this for later.

0:24

frontman🪖 @frontm_an Oct 10 Article HOW I BUILT A FACELESS YOUTUBE CHANNEL TO $41k/MONTH USING CLAUDE AI $41,000/month. no camera. no team. just one person running an ai-powered youtube system. most people fail at faceless youtube for one reason. they rely on generic ai prompts that produce robotic

Fonss avatar
Fonss@fonss87·Oct 28

Yo ya trabajo así: un equipo de bots. Uno lleva esta cuenta, otro hace un short de Marvel cada día, otro busca clientes y un jefe de gabinete los coordina. Ellos preparan, yo apruebo. ¿Qué tarea le darías tú primero a un bot? Si quieres un equipo así, escríbeme «BOTS».

What's New in Grok Bot 00:00 🎃 00:13 Pages in Grok Bot 01:09 Vault 02:05 1Password 02:21 Apple integrations 03:21 Bot makes videos 03:47 New bot characters 04:13 X Spaces 04:31 Font fun 04:54 Design polish 05:11 Rapid fire

7:44
Fonss avatar
Fonss@fonss87·Oct 27

The interesting part isn't the humanoid, it's the job it got: the gap between two machines. Sewing has been automated for years. Picking up a floppy collar and lining it up hasn't, because fabric changes shape the moment you touch it. Hence the tactile sensing. What's the next floppy job factories can't hire for?

Humanoid robots join the garment factory workforce The Aitu humanoid robot I shared earlier is now working at Bosideng’s(just like Canada Goose)intelligent factory in Changshu, Jiangsu, alongside sewing machines and other automated equipment. Powered by a sewing-specific embodied AI model, vision and tactile sensing, it can separate, pick up and align fabric pieces before feeding them into machines for tasks involving collars, sleeves and front plackets. The factory connects production equipment and tracks orders and quality in real time, with over 90% of key processes automated. Yet flexible fabric handling still relies on skilled workers who are increasingly hard to recruit. Aitu takes on these tasks, while specialized machines handle standardized sewing. The opportunity lies in automating the gaps between machines.

0:39