An overview of interesting AI News for the Yate and Sodbury U3A's monthly AI News meeting held on March 12th 2026 (meetings are second Thursday of month, 1400-1500, St John's Church Centre, Wickwar Road, Chipping Sodbury, UK - all welcome). Note that this is a summary to act as background as the topics are discussed in detail at the meeting itself. We sometimes have extended meetings where we demonstrate other AI products and services.
The 'Introduction to AI' course we run in Chipping Sodbury now uses these News sessions to replace the original classes 5 and 6 as the topics in them (humanoid robots, self driving cars, ethics, futures) are now 'news' not 'distant futures'.
In the film Idiocracy people evolved to become less smart. In the Matrix people were plugged into a virtual world. In the Terminator the robots might kill you. Elon Musk responded "100%" to this diagram.
— Rothmus 🏴 (@Rothmus) March 8, 2026
Remembered particularly for the creative 'Move 37' and what it implied for AI's future.
Ten years ago, AlphaGo’s legendary match in Seoul heralded the start of the modern era in AI. Its famous ‘Move 37’ signaled to us that AI techniques were ready to tackle real-world problems in areas like science - and ideas inspired by these methods are critical to building AGI pic.twitter.com/8EibfAByaG
— Demis Hassabis (@demishassabis) March 10, 2026
One of the most interesting measures of AI capability is the METR series which basically tracks AI's ability to complete typical human tasks by the length of time humans take to to achieve them. At the moment the best AI might manage a 14 hour task. Now see what happens after a year if that period continues to double every 129 days!
We're correcting a mistake in our modeling that inflated recent 50%-time horizons by 10-20% (and reduced 80%-horizons). We inappropriately penalized steepness in task-length→success curve fits. This most affects the oldest and newest models, whose fits are less data-constrained. pic.twitter.com/YKbuLEwG9n
— METR (@METR_Evals) March 3, 2026
How 'AI Agents' completely change the economics of writing software by both speeding up development and drastically reducing cost.
"Not a single one of our full-time engineers right now writes code anymore... it's all agentic... the way you build a company completely changes." @justindross with @JTLonsdale on the future of software:
— American Optimist (@AmOptimistShow) March 2, 2026
JD Ross: Not a single one of our full-time engineers right now writes… https://t.co/TJ8PiIwKV2 pic.twitter.com/o97Td9q8IT
Speed comes at a price though, in terms of reduced accuracy. The demonstration remains very impressive though, and the improvement over the previous Gemini Flash model is huge in both speed of completion and accuracy.
Gemini 3.1 Flash-Lite has landed.
— Google DeepMind (@GoogleDeepMind) March 3, 2026
It’s our most cost-efficient Gemini 3 series model yet, built for intelligence at scale. Here’s what’s new 🧵 pic.twitter.com/BzD2bdg3Dx
April is the start date for volume production and observers have seen 25 Cybercab vehicles around the factory, being used for testing and crash safety tests. Meanwhile there are a limited number in the field trials. Will Tesla be able to bring very large numbers into use when they come off the production line?
Big day for Cybercab at Giga Texas today! Actually, yesterday to kick off March, the production line went into a higher volume & today we see 25 at three main locations, and there were several others I observed driving around too!
— Joe Tegtmeyer 🚀 🤠🛸😎 (@JoeTegtmeyer) March 3, 2026
I think this may be the largest single grouping… pic.twitter.com/HZDMNv57lJ
You will be able to upgrade your robotic vacuum cleaner to a fully trained humanoid cleaner if Figure succeed in their mission, just one sector they are targetting.
Today, Figure is showing another major milestone towards a robot in every home
— Brett Adcock (@adcock_brett) March 9, 2026
Running Helix 02, cleaning a living room fully autonomously pic.twitter.com/tKIkiDDuCR
"The concept of owning the full stack on an infinite labour machine that can rebuild the earth and essentially create new industries from the ground up is probably the biggest theoretical win that any company could have ..." Chris Camillo
Investor Chris Camillo explains the real $TSLA endgame: infinite labor 👀
— The Iced Coffee Hour (@TheICHpodcast) February 28, 2026
“An infinite labor machine that can rebuild the earth and create new industries from the ground up is probably the biggest theoretical win any company could have.
If generalized robotics materializes over… pic.twitter.com/hequDt8IOg
Apparently "Hey Plex" are the magic words to summon AI Perplexity to help you with whatever you desire.
Perplexity is now integrated into all Samsung Galaxy S26 phones. Every S26 device will have Perplexity pre-loaded with the wake word "Hey Plex". The devices come with three assistants: Perplexity, Bixby, Gemini. Bixby will also be powered by Perplexity's search-grounded LLMs. pic.twitter.com/LM9OwkGxry
— Computer (@AravSrinivas) February 26, 2026
Anthropic wished to impose terms that prevented the use of Claude in autonomous drones etc (i.e. no human in the loop situations) and also use for surveillance of civilian populations. Whereas this sounds sensible, those engaged in preparing for warfare hate the idea of having to consult lawyers. Meanwhile OpenAI and X seem to have reached agreement, while Anthropic has seen many cancelled government contracts (and are taking legal action about that).
A statement from Anthropic CEO, Dario Amodei, on our discussions with the Department of War.https://t.co/rM77LJejuk
— Anthropic (@AnthropicAI) February 26, 2026
Ever wanted to make an image that reflects the actual weather in your chosen location? Probably not, but that and other advanced features are now available with Nano Banana 2.
Introducing Nano Banana 2, our best image model yet 🍌🍌
— Sundar Pichai (@sundarpichai) February 26, 2026
It uses Gemini’s understanding of the world and is powered by real-time information and images from web search. That means it can better reflect real-world conditions in high-fidelity.
Check out "Window Seat," a demo… pic.twitter.com/zhe8tC4jCL
There are many auto manufacturers who see that they will need to offer autonymous driving but need someone to do the hard work for them. Wayve offers a potential solution, though the obstacles to integrate it to existing models remain tremendous.
“It’s clear autonomy is now coming to market in a massive way.”
— Wayve (@wayve_ai) February 25, 2026
Wayve CEO @alexgkendall joined @business this morning to discuss our $1.5B fundraising announcement.
It's clear validation of Wayve as a global player in the AV industry, with backing by global automakers… pic.twitter.com/mILQI5365J
There are many reports of AI tools making substantial progress in solving mathematics problems. This one seems significant because of the high reputation of the mathematician crediting Claude.
Legendary mathematician Donald Knuth reveals Opus 4.6 solved his long-standing conjecture:
— Haider. (@slow_developer) March 4, 2026
"claude opus 4.6 cracked my long-standing hamiltonian-cycle conjecture for all odd sizes — an open problem from my art of computer programming drafts, and it's "a joy" to see it solved"… pic.twitter.com/tG2WOLdP0a
This was just one of a number of significant falls in the prices of shares of well known software companies as Anthropic released tools for Claude that solved previously difficult problems in software development, in this case rewriting COBOL based software.
BREAKING: IBM stock, $IBM, falls over -10% after Anthropic announces that Claude can streamline COBOL code.
— The Kobeissi Letter (@KobeissiLetter) February 23, 2026
It’s becoming increasingly clear how pivotal the times we are in right now truly are. pic.twitter.com/yGVZBeHk3R
These are their findings in a more readable style than the one they used originally.
A universal truth: most radar charts should just be bar charts.
— Peter Walker (@PeterJ_Walker) March 6, 2026
Love your stuff, Anthropic! pic.twitter.com/DvWbXpAe3w
They just launched The Anthropic Institute. "... society is shortly going to need to confront many massive challenges. How will powerful AI systems reshape our jobs and economies? ... What kinds of threats will they magnify or introduce?"
Introducing The Anthropic Institute, a new effort to advance the public conversation about powerful AI.https://t.co/M7vi9oRuYi
— Anthropic (@AnthropicAI) March 11, 2026
AI researcher Connor Leahy discusses the perils of our rapid progress in building AI
“We built something we can't control."
— Peter McCormack 🏴☠️🇬🇧🇮🇪 (@PeterMcCormack) March 10, 2026
AI researcher Connor Leahy, @NPCollapse, explains why the people building super-intelligence are flying blind.
We cover:
🚨 Inside the "Black Box"
🚨 The rise of "AI Psychosis"
🚨 Why AI chooses nukes
Link to full episode 👇🏼 pic.twitter.com/5lvTB1Kt2F
New York Times writer Exra Klein says we have to switch from "What Happens If" to "What Happens Now"
Well said by Ezra Klein in the NYT this morning.
— Nathan Calvin (@_NathanCalvin) March 9, 2026
"For years now, questions about A.I. have taken the form of "what happens if?"
. . .
This year, the A.I. questions have taken a new form, "what happens now?" pic.twitter.com/Es2gvGRRGX
Runway Characters combine an intelligent AI with a live character avatar. The BBC is using one in a show, though they have carefully avoided anything too lifelike.
Last week we debuted our new real-time video agents with one of the hardest demos possible: live television. The BBC is now using Runway Characters to augment segments of their programming. Wild to see this live. So excited for all the new possible applications to come. pic.twitter.com/E0JgOYx33Z
— Cristóbal Valenzuela (@c_valenzuelab) March 9, 2026
Most AI tools of the future will run on small local devices, of which this is an early extreme example.
Can’t believe they actually shipped it 😂
— Bo Wang (@BoWang87) March 10, 2026
A $20, bottlecap-sized 🦞 in your pocket 🔥 https://t.co/TwD21q4C2j pic.twitter.com/stY9fe8gga
Before long it wil be difficult to tell who is human and who is a humanoid
It was only a matter of time before humanoid robots would get human faces.
— Chubby♨️ (@kimmonismus) March 9, 2026
AheadFrom Robotics getting less uncanny. pic.twitter.com/atf5pRCqge