A week in Generative AI: Robot Special!
News for the week ending 23rd February 2025
Compared to the last few weeks, this week has definitely been a much quieter week on the generative AI front. However, we’re definitely in for a busy couple of weeks as we close out February - OpenAI are likely to be launching GPT-4.5 next week and it looks increasingly likely that Anthropic will be releasing their next model in the next week or so.
But not to worry, there happened to be quite a lot of fun robotics news this week, so below are some great robot videos to tide you over! 🤖
In ethics news, OpenAI revealed that they now have 400M active users every week, which is a big increase on the 300M last reported and NVIDIA’s CEO spoke about how the markets got it wrong with DeepSeek.
In long reads I highly recommend Benedict Evan’s post on The Deep Research problem and watching the Economist’s interview with Demis Hassabis and Dario Amodei.
Thinking Machines Lab is ex-OpenAI CTO Mira Murati’s new startup
Mira Murati left her role as CTO at OpenAI back in September to do her ‘own exploration’. This week she announced what that is, a new company called Thinking Machines that she’s set up with OpenAI co-founder John Schulman and OpenAI’s ex-chief research officer Barret Zoph.
Thinking Machines will focus on “key gaps” that they see in current AI capabilities. They plan to build multimodal systems that can work with people collaboratively and enable a broader spectrum of applications than are currently available. The initial focus will be on science and programming applications as they want to enable novel scientific discoveries and engineering breakthroughs.
This all sounds very ambitious and interesting, and they’ve definitely got the leadership and talent to make an impact. Looking forward to what they come up with in the coming months!
Norway’s 1X is building a humanoid robot for the home
We are definitely going to be seeing more robots in our homes over the coming decade and it’s likely that some people have a humanoid robot of some description to help them. This video is a vision of what that could be like from Norway’s 1X who have previously released some impressive robotics videos.
I do find this video a bit strange though - the robot is seen vacuuming, which we already have specialised robots for, and even taking a break on a sofa at the end of the video! Fun to watch though.
Figure’s humanoid robot takes voice orders to help around the house
This latest update from Figure shows the new proprietary AI model powering their robots. They used to have a partnership with OpenAI to provide the onboard smarts for their robots but recently decided that they needed specialised models and so have gone their own way.
The video is impressive - it shows the robots reasoning through tasks and putting items away without being told where they should go. The most impressive thing however is seeing two humanoid robots working together - that’s not something I’ve seen before and I’m interested in whether there is some communication going on between them or if they’re just reacting to the body language of the other robot.
Israeli firm unveils humanoid robot with 360° vision, 3x power and precision
Mentee have been doing some great robotics work over the last couple of years, and this is their new platform - MenteeBot v3.0. They’ve got 360 degree cameras and a swappable battery so that they can carry on working without taking a break to charge.
Protoclone
This is both incredibly impressive engineering and incredibly freaky to see all at the same time. Clone Robotics are a company that is taking a completely different approach to most robotics companies right now - they have been doing a lot of R&D into musculoskeletal robotics, essentially robotics that are powered by artificial muscles, much like how the human body works. The result is a much higher degree of movement that traditional robotics and much more strength - they claim their robots are 10x stronger than existing alternatives.
I have no doubt that this is the future of robotics, but at this research phase it’s very unnerving to see as it is straying into the uncanny valley!
AI Ethics News
NVIDIA CEO Jensen Huang says market got it wrong about DeepSeek’s impact
AI cracks superbug problem in two days that took scientists years
Long Reads
Benedict Evans - The Deep Research problem
The Economist - AI bosses on what keeps them up at night
Dwarkesh Patel - Satya Nadella ơn Microsoft’s AGI Plan & Quantum Breakthrough
“The future is already here, it’s just not evenly distributed.“
William Gibson
I don't see this as an either/or - ideally we need both to move the technology forward as much as possible, whilst also ensuring as many people can access the technology as cheaply as possible.
Do you think that companies should focus on more realistic tech products rather than products such as humanoid robots that most likely will only be available to a small percentage of people. In my opinion the development of technology is good however I think everyone benefits more from more accessible technology.