AI News - October 2025

An overview of interesting AI News for the Yate and Sodbury U3A's monthly AI News meeting held on October 9th 2025 (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 'futures'.

[1] OpenAI's Sora 2 provides instant social media clips

The original Sora led the way in 'prompt to video' AI back in early 2024. Many others have competitors have emerged since but with this new release OpenAI are back in the lead (for this month anyway, possible Google Veo 3 upgrade coming). Sora 2 take any prompt and refines it to produce a video clip with many interesting features including dialogue or music and lip-synched action. This is then wrapped in a new social media site concept.

[1.2] In this example the creator says that they did not provide any of the dialogue, just the situation.

[1.3] And in this example the prompt was as simple as "In South Park the gang discovers Waymo". Sora 2 allows use of copyright material unless the owner opts out.

[1.4] You can also create a personal 'cameo' video and then allow anybody else to include your character in your creation, as Sam Altman has done. This can then be placed on the inbuilt Sora 2 social media site. The hope is that making clips combining famous people and characters with your own image and that of friends will be a social media hit.

It is also true that Sora 2 could provide a literally endless stream of entertaining video clips all by itself. If you are not a creator yourself (i.e. a prompt suggestion person) there would be no detectable difference. Humans will soon be relegated to a role of just watching.


[2] Sam Altman on AI and the public

Sam discussing how most people are still using only a small part of the available AI capability, and the implications of that.

Sam discusses how "this was the year that the AI got smarter than us". Also refers to TV Show Pantheon.

Pantheon explores the concept of Uploaded Intelligence (UI), where human consciousness can be digitized and uploaded into the cloud, raising profound ethical, societal, and technological questions.


[3] A look back at last year's Nobel Prize

There are more serious uses of AI underway. Last year in chemistry the award was to Demis Hassabis and John Jumper for using AI to predict the structure of almost all known proteins. The diagram below is a fascinating explanation of how that worked if you have time to look at it.


[4.1] Old monkeys made younger

Or as the scientific paper puts it: 'Senescence-resistant human mesenchymal progenitor cells counter aging in primates' with results like restored memory, reversed brain shrinkage and bone loss, AND rejuvenated half of all tissues.

[4.2] Bryan Johnson is a high profile figure in the Longevity arena. He suggest maybe this could give humans 9 to 15 years of rejuvenation.


[5] The makers of Perplexity AI make their new AI browser Comet available to all

This has been around for a while but only for paid subscribers. Now they have put it on general release including for FREE users. This is a major step towards a future where you do everything through a single interface.


[6.1] Tesla Optimus Humanoid Robot learns Kung Fu

Elon confirmed that this was NOT tele-operated. The Optimus AI brain is making the moves.

[6.2] Tesla Optimus showing celebrity potential.

This is the launch of Tron where the Kung Fu was on show again. Soon humanoids will be an essential part of most any major publicity event.

[6.3] Tesla Robotaxi growing fast.

Now covering more cities and gaining market share with low pricing. Everyone waits to see when the supervisors in the passenger seat will be removed.

[6.4] Tesla makes two significant car related announcements this week.

This week Tesla released new lower cost versions of their top selling Model 3 and Model Y cars, and also Full Self Driving v14, a very major release. Here is a short journey using the FSD version.


[7.1] Massive Drone Swarm from one shipping container ...

The person posting points out that the 720 drones in one container might be enough to take on the entire tank forces of some western armies.

[7.2] ... but Drone swarms remain vulnerable to jamming.

[8] The Doordash Home Delivery robot.

Carries up to 6 pizzas apparently.

[9] OpenAI lets you buy within Chat.

A massive income stream incoming for OpenAI as you can first let ChatGPT find you a product or service then seamlessly order it for you. Another reason why most ordinary people will get powerful AI for free.

[10.1] Meta have added to their AI Smart Glasses range with an 'in-vision' display screen version

Lots of applications including for deaf people and for language translation. Here used as a navigation aid.

[10.2] Apple to focus on AI smart glasses (not headsets).

Following the success of Meta AI smart glasses, and the relative failure of the Apple Vision Pro headset, it seems likely Apple will soon have an AI Smart Glasses range.


[11] Marc Andreessen on the usefulness of AI.

Making the point that everybody has access to powerful AI and discusses 'AI reluctant' people.

[12] Richard Sutton talking about the inevitability of what is happening with AI.

In particular that we won't stop with human level intelligence, and the implications of that.


[14] Xai's Grok (Twitter) continues rapid progress

The prompt to Image to Video tool called 'Imagine' continues to improve. Gives lots of free use so worth a look.


[15] Neuralink brain implant controls robotic arm

The number of implants grows as do the uses of the mind computer control


[16] Mattel use Sora prompt to video to design product

It is easy to see how the design process can be much faster if a sketch can turn into a visualisation in minutes