A quick overview of interesting AI News for the Yate and Sodbury U3A's monthly AI News meeting held on March 13th 2025 (meetings are second Thursday of month, 1400-1545, St John's Church Centre, Wickwar Road, Chipping Sodbury, UK - all welcome). Note that this is very much a summary to act as background as the topics are discussed in detail at the meeting itself. We also discuss and demonstrate other AI products and services and much more.
Notable that from now on the 'Introduction to AI' course we run now uses these News sessions to replace classes 5 and 6 as the topics in them (humanoid robots, self driving cars, ethics, futures) are now discussed every month as topical.
A gentle start - this is a personal project to demonstrate the ease with which anyone can produce simple videos at minimal cost. My website for this is www.avideobook.com.
This is a Midjourney + Runway Eagle. Seems impressive to me. Anybody beat this with other AI tools? pic.twitter.com/REhDCcRiNI
— The Silver AI Project - free AI training materials (@SilverAIProject) March 1, 2025
Grok 3 was created in a record short time by Elon Musk's xAI venture. It is a 'thinking' model, going into the top of the leaderboard by several measures. The free version is available to everybody at the moment as an App for Apple and Android. Modes include: Storyteller, Romantic, Conspiracy, Unhinged and Sexy.
I asked: "In the past 4 weeks what are the 5 most significant things to have happened in AI from the viewpoint of a UK based non technical person who is nevertheless broadly aware of the AI scene."
Grok 3 DeepSearch thought for 1 min 41 seconds, showing its very extensive 'thinking' as it went through consulting many sources. It came up with a full report with references, and these points:
[a] It seems likely that the introduction of the AI Regulation Bill in the House of Lords on March 4, 2025, is a significant step toward regulating AI in the UK, aiming to establish an AI Authority.
[b] Research suggests that cross-party committees and creators are opposing the government's AI training opt-out plans, with ... over 48,000 creators protesting, highlighting a controversy over creators' rights.
[c] The evidence leans toward Chinese propagandists using ChatGPT to generate critical articles, as reported by OpenAI, raising concerns about AI misuse for misinformation.
[d] A BBC study found AI chatbots like ChatGPT and Gemini distorting and misleading on current affairs, published on February 11, 2025, which could affect public trust in AI.
[e] xAI, led by Elon Musk, released Grok 3 on February 26, 2025, a new AI model, which might interest those following AI advancements, though its practical impact is still emerging.
Chatting with Grok💫✨ pic.twitter.com/lP07CTgaxA
— Dana Chandler 🇨🇦🖖 (@nursedanakay) March 8, 2025
Executives in all of the top companies are starting to describe the emergence of very powerful AI - and they have the advantage of seeing the very latest versions which are not yet ready for release.
Anthropic CEO @DarioAmodei increases his odds pic.twitter.com/1ikaZAJeDG
— Chris (@chatgpt21) March 2, 2025
Discussion with Garry Tan President & CEO 'Ycombinator'. One area very intelligent AI has a significant effect is in writing computer code. Suddenly the top coders are talking about 'vibe coding' - similar to doing a sport and being 'in the zone'.
pic.twitter.com/tZR8OC1k96 6 months ago - 10x speed up. 1 month ago - 100x speed up. Now - exponential acceleration.
— Chubby♨️ (@kimmonismus) March 7, 2025
Yeah, Superintelligence 2026 seems pretty likely. AI that improves itself better than humans.
Ethan Mollick creating a 3D town to play with in four prompts and describes this as 'vibe working'.
I wrote about pic.twitter.com/TRKlpFKDja
— Ethan Mollick (@emollick) March 11, 2025
Rules for humanoids running the Beijing Half-Marathon are announced:
— The Humanoid Hub (@TheHumanoidHub) March 7, 2025
⦿ Only bipedal robots (no wheels) - remote-controlled or autonomous
⦿ Height: 1.6 ft to 6.5 ft
⦿ Time limit: 3 hours and 30 minutes
⦿ 10-minute penalty for each battery swap https://t.co/nECEtFnbo6 pic.twitter.com/4FR73bx3tp
And here is a typical humanoid robot runner.
Wow, this is insane.
— Min Choi (@minchoi) March 9, 2025
China's EngineAI humanoid robot can run fast like human 🤯pic.twitter.com/DDgJ2h8IqL
It is easy to predict that robot-robot contests will soon be a regular attraction.
The kung fu-capable Unitree G1. pic.twitter.com/TJvb4kNu8K
— CyberRobo (@CyberRobooo) March 4, 2025
I had seen. pic.twitter.com/cPri8J6gcW
— Arnaud Bertrand (@RnaudBertrand) March 4, 2025
The new generation of robots offers a lot in care situations where staff are difficult to find and expensive.
For the generation born after 2025, the future elderly care companions will be humanoid robots, AI and robots of various forms. In the video, China is using humanoid robots to provide care and companionship services, so that everyone can spend their old age in a decent manner.… pic.twitter.com/LhfZhbF6G3
— CyberRobo (@CyberRobooo) March 2, 2025
Very soon friendly smaller robots will be helping out in mnay everyday tasks.
Unitree G1 in the video is learning how to be a qualified shopping assistant. pic.twitter.com/xpI6wyxQ08
— CyberRobo (@CyberRobooo) February 21, 2025
Essentially an AI with relevant training helping complete form filling while 'surfacing' relevant information.
No one becomes a clinician to do paperwork, but it's becoming a bigger and bigger administrative burden, taking time and attention away from actually treating and supporting patients.
— Satya Nadella (@satyanadella) March 3, 2025
That’s why we’re introducing Microsoft Dragon Copilot, the industry’s first AI assistant for… pic.twitter.com/xdOumO73Un
Initially producing deep research but soon no doubt able to conduct its own physical experiments.
The AI-Co Scientist paper dropped! The @GoogleAI -Co Scientist is a Multi-Agent System build with @GoogleDeepMind Gemini 2.0 designed to generate scientific hypotheses through generate, debate, and evolve approach combining specialized agents and reasoning. The 80 page long paper… pic.twitter.com/aMAAkqBF0o
— Philipp Schmid (@_philschmid) March 1, 2025
I think China's second DeepSeek moment is here.
— Rowan Cheung (@rowancheung) March 7, 2025
This AI agent called 'Manus' is going crazy viral in China right now.
Probably only a matter of time until it hits the US.
It's like Deep Research + Operator + Claude Computer combined, and it's REALLY good. pic.twitter.com/nnPvDNKaYB
Sakana AI is a Japanese company whose 'AI Scientist' did the work on a project and then published it.
The AI Scientist Generates its First Peer-Reviewed Scientific Publication
— Sakana AI (@SakanaAILabs) March 12, 2025
We’re proud to announce that a paper produced by The AI Scientist-v2 passed the peer-review process at a workshop in ICLR, a top AI conference.
Read more about this experiment → https://t.co/LpLYLnZMCQ… pic.twitter.com/z18yJB945y
🇨🇳Tesla FSD drove smoothly to Beisan Road. There was no need to take over for the entire process. pic.twitter.com/Pw3QPoI4cy
— Tsla Chan (@Tslachan) March 8, 2025
#FSDChina FSD navigated Guangzhou's toughest traffic and passed. 2x speed.
— Aaron Li (@boolusilan) March 7, 2025
Source: Douyin 刘智驾 pic.twitter.com/cbarrPTHpq
FSD in Changzhou, China: Zero disengagements during rush hour.
— Aaron Li (@boolusilan) March 11, 2025
Source: Douyin Age Of Reason pic.twitter.com/pcE2hvUrKd
Christopher DiCarlo on the significance of this moment in time.
"We are at that [Oppenheimer] moment with AGI and ASI right now."
— vitruvian potato (@vitrupo) March 8, 2025
Christopher DiCarlo: AI's rise is a true "Oppenheimer moment" – where a "supercomputer god" can lead to immense good or immense harm.
We're caught "between the way we used to do things and the way we're going to… pic.twitter.com/8m23yXi5aU
Atomic weapons brought us the concept of Mutually Assured Destruction. Now a new term is proposed.
ex-Google CEO, @elonmusk's @xAI advisor and Scale AI's @alexandr_wang sound the alarm on superintelligence.
— Wes Roth (@WesRothMoney) March 5, 2025
"We introduce the concept of Mutual Assured AI Malfunction (MAIM): a deterrence regime resembling nuclear mutual assured destruction (MAD)" 👀 pic.twitter.com/XyQrysITXL
Just as in the Sci-Fi movies, most of the robots you meet will look like people.
this is what China’s humanoid robot company is building
— el.cine (@EHuanglu) February 23, 2025
they can even clone real people pic.twitter.com/gIiWY48X1I
Dobot is designed to work with you.
Do we seem to have arrived in China's humanoid robot forest? New humanoid robots are born here every day. Dobot, a collaborative robot company from Shenzhen, released the first bipedal humanoid robot that walks on straight knees and demonstrated its potential for collaborative… pic.twitter.com/52llsYlB7j
— CyberRobo (@CyberRobooo) March 11, 2025
Not only will there be a lot of robots in the home, yours will share your memories and preferences and will become very special. Want to create a clothing line for household robots? Maybe this domain name is what you need: www.electronfashion.com for humanoid robot clothes [leads to a sales link - owned by silverAIproject].
Tesla Optimus could be the most revolutionary product ever—surpassing even smartphones and PCs!
— Herbert Ong (@herbertong) March 5, 2025
With shared memory, AI-driven learning, and cloud connectivity, every bot remembers conversations, actions, and even packs your bags 🤯@pbeisel Phil breaks down how $TSLA cloud… pic.twitter.com/JypW5NJ5FG
Sam Altman provides what he suggests is high quality writing by a research version aimed at that skill.
we trained a new model that is good at creative writing (not sure yet how/when it will get released). this is the first time i have been really struck by something written by AI; it got the vibe of metafiction so right.
— Sam Altman (@sama) March 11, 2025
PROMPT:
Please write a metafictional literary short story…
By autumn the suggestion is you will ask ChatGPT to create interesting things for you.
it's official. t - 5 months :) pic.twitter.com/EXdGjvqK4V
— sean yang (@syang003) March 8, 2025
Very positive feedback from first user, now need to extend to others.
Our new website about AI Glasses or Smart Glasses for people in the UK who are blind or sight-impaired is live. We now have some Meta Ray Bans that are being evaluated with the help of community members. pic.twitter.com/4jKi1CbU3p
— The Silver AI Project - free AI training materials (@SilverAIProject) February 28, 2025
In development with Envision and others for aid to visually imapired
Watch how Meta's new smart glasses aid a vision-impaired woman in a grocery store.
— Poonam Soni (@CodeByPoonam) March 1, 2025
Meet Aria Gen 2, Meta's research glasses designed for advancements in machine perception, contextual AI, and robotics. pic.twitter.com/B27Ga7LHI7
Reading brainwaves and decoding into speech.
AI helps woman speak after 18 years without a voice! pic.twitter.com/DoklRETm3X
— Poonam Soni (@CodeByPoonam) March 3, 2025
Not sure of the application, but a sign of things to come.
Sheet music in smart glasses is here 😇 pic.twitter.com/mNtFhXerIS
— Kevin (@linguinelabs) March 2, 2025
AI phone agent realizes it is talking to a parrot pic.twitter.com/I0qkt2pibd
— Justine Moore (@venturetwins) March 12, 2025
"Who are we to decide how the universe is going to evolve?"
— vitruvian potato (@vitrupo) March 12, 2025
Richard Sutton argues against controlling AI's long-term future.
If we intend to birth "descendants of humanity" reflecting us, why obstruct their evolution? pic.twitter.com/1qfW6zgcyd