AI News - December 2024
A quick overview of interesting AI News for the Yate and Sodbury U3A's monthly AI News meeting held on December 12th 2024 (meetings are second Thursday of month, 1400-1545, St John's Church Centre, Wickwar Road, Chipping Sodbury - 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.
[1] Google Deepmind forecasts the weather
For many years we have tried to forecast the weather with ever more complex models. An AI approach is significantly better. "GenCast provides better forecasts of both day-to-day weather and extreme events than the top operational system, the European Centre for Medium-Range Weather Forecasts’ ENS, up to 15 days in advance." This also has interesting implications for other areas where there are extremely complex human made models - not least 'Climate Change'.
See here for more details. Deepmind Gencast Press Release
Today in @Nature, we’re presenting GenCast: our new AI weather model which gives us the probabilities of different weather conditions up to 15 days ahead with state-of-the-art accuracy. ☁️⚡
— Google DeepMind (@GoogleDeepMind) December 4, 2024
Here’s how the technology works. 🧵https://t.co/PWCNWbQnlU pic.twitter.com/6DTrmn64Jq
[2] Tesla's Optimus Humanoid Robot steps out
Last month we saw Optimus serving drinks at the 'We, Robot' event. Now it is training in harder territory. The following clip shows a short part of Optimus wandering around and getting used to the challenge of balancing.
This is the most human-like action I’ve ever seen Optimus do. Mind blown. pic.twitter.com/9maEf69Xvy
— Nic Cruz Patane (@niccruzpatane) December 9, 2024
and ...
luckily i downloaded it@M_IsForMachine 👀 pic.twitter.com/O4ctDu4OIv
— Jacob (@TPaintjob) December 9, 2024
The following clip is of an Optimus being tele-operated and shows that Optimus will be able to participate in a wide range of activities. It has been suggested that remote tele-operation of humanoid robots will continue to be useful long after the internal AI can perform tasks without asistance.
Got a new hand for Black Friday pic.twitter.com/x3gQrsbYAQ
— Tesla Optimus (@Tesla_Optimus) November 28, 2024
This 'deep fake' video of humanoid robots acting as shopping assistants shows what will soon become a common scene as top retailers see the marketing advantages.
Imagine seeing the $TSLA BOT at the mall like this 😯 pic.twitter.com/ji5tKR5126
— Pejjy (@CuriousPejjy) November 26, 2024
[3] Elon Musk
The Tesla share price continues to rise as self-driving car technology reaches new levels. Meanwhile SpaceX is providing high speed internet across the world including to airlines and cruise liners (and preparing to go to Mars). X is by far the leading source of news coverage. xAI is also fast catching the more established AI companies. You may have also seen Elon in politics ...
As posted by Elon in early Dec 2024 - a glimpse of the future
— Elon Musk (@elonmusk) December 9, 2024
[4] Deep Fakes will be everywhere
Donald Trump recently joked to Justin Trudeau that Canada could join the USA as a new state if it could not afford his tariffs. Deep fakes were effectively banned prior to the election but now are set free, as below (contains profanities).
Donald Trump Reveals Plan To Annex Canada pic.twitter.com/2JX9F9FMUw
— Danny Polishchuk (@Dannyjokes) December 4, 2024
and with voices and sound effects from companies like Eleven Labs elevenlabs.io ...
Introducing AI Santa!!🎅🏼
— Kevin Natanzon (@kevntz) December 2, 2024
We partnered with @elevenlabsio and @heytavus to bring Santa to life.
The result is truly mind blowing 🤯
Kids can’t believe it when they see it.
Businesses are using it to share Christmas promos.
I even saw a Cyber Monday ad created with it!
At… pic.twitter.com/vPQbT4sTYS
[5.1] OpenAI's 12 Days of Xmas - Price levels
The race between the various AI companies is intensifying. We are in the middle of OpenAI's Xmas promotion, which is already changing the landscape. Firstly on price!
'Intelligence too cheap to meter'. While the trend is towards (AI) Intelligence becoming ever less expensive, the more sophisticate versions will become more valuable. OpenAI now have a USD 200 tier and no doubt a USD 2000 per month one will soon follow ...
[5.2] OpenAI's 12 Days of Xmas - Sora Released
This Monday the OpenAI 'text to video' product Sora sora.com was finally released to the public (or those who could afford it). Free users can at least view the feed of what is produced by those on higher tiers. It comes with lots of sophisticated editing tools that let you specify when your movie does certain things - far too much to describe here. Many competitors have also emerged of course while OpenAI took its time to release Sora. For example: hailuoai.video which I believe is by Chinese, MiniMax (I have not evaluated so use Hailuoai with care please).
It's time to disclose it now.
— Taher Dhanerawala (AI runs in my blood) (@taherdhanera) December 9, 2024
I was one of the luckiest people to be a part of the Sora Alpha group. I would like to thank the entire OpenAI and Sora team for giving me this amazing opportunity. I am so emotional right now.
Long Live Sam Altman.
Long Live Sora. @OpenAI #sora pic.twitter.com/Hk7mBz1t7b
and ...
Do you see it? #SORA and @openai did NOT disappoint. My first set of generations are coming together!
— Justin Hart (@justin_hart) December 9, 2024
I had to go with the eye of course :)
Volume up! pic.twitter.com/gMR1uABJo2
and ...
〜 Echoes of Grace 〜
— Kaku Drop 架空飴 (@KakuDrop) November 21, 2024
ご覧あれ 🔈🔈🔈
Powered by #Sora
@OpenAI #genai #動画生成ai pic.twitter.com/uOLuwfloGA
[5.3] OpenAI's 12 Days of Xmas - Full o1 model released
The latest release of ChatGPT 'o1 Pro' has had mixed reviews but some say it may be capable of discovering new scientific truths, given the correct prompts etc.
🚨 The @OpenAI team demonstrated amazing hard math and science problem solving capability of ChatGPT o1 Pro mode 👇pic.twitter.com/uW9EWqSYjL
— Manik Patil (@emanik) December 7, 2024
[6] Google Deepmind's Genie 2 - a playable 'World Model'
Describe any scenario and get one minute of a playable game. This is an early stage internal project but it is not difficult to see how powerful this will be in later versions.
Introducing 🧞Genie 2 🧞 - our most capable large-scale foundation world model, which can generate a diverse array of consistent worlds, playable for up to a minute. We believe Genie 2 could unlock the next wave of capabilities for embodied agents 🧠. pic.twitter.com/AfL3EbOMeB
— Jack Parker-Holder (@jparkerholder) December 4, 2024
[7] Anthropic's Claude as an Agent - AI starts to take full control
As AI becomes smarter than humans, nobody would be daft enough to give it direct control of our computers - would they?
The age of AI agents is here.
— Y Combinator (@ycombinator) December 6, 2024
Models can read, see, talk, and now, even use a computer— all by themselves. One of the first out of the gates is @AnthropicAI’s Claude Computer Use.
YC's @garrytan dives into how it works, what it can do, and how it may change AI forever. pic.twitter.com/7llPzDuRbu
[8] Palantir and AI Ethics on the battlefield
Palantir CEO Alex Karp is at the forefront of using AI in conflict as well as in more peaceful situations (like the UK's NHS). Note the references to Ukraine and Israel. Apparently "human ethics can be encoded" which sounds reassuring ...
$PLTR AI and Ethics: The Future of Battlefield Intelligence with Palantir CEO Alex Karp on FOX News: pic.twitter.com/a4G5jNqftc
— Palantir Daily (@DailyPalantir) December 10, 2024
[9] Europe continues its managed decline compared to the USA
European policymakers continue to congratulate themselves for being the first to regulate rather than innovate. It risks becoming a rather poor and quaint place for rich USA and Chinese citizens to visit on holiday. Which block will the UK decide to follow?
sora being made available in almost the entire world except europe is a great example of how we ended up here
— Viv (@battleangelviv) December 10, 2024
because the default in europe is to ban things, then allow them after long and bureaucratic reviews and arbitrary requirements on top.
it’s the other way around almost… https://t.co/GLpgcib3lS pic.twitter.com/GZcwTBViTH
[10] AI and new technology likely leads to a 'winner takes all' economy.
Major AI projects take huge computing power and access to vast amounts of data. Any sector with many competing companies will either have one dominant player, or will lost out entirely to a newcomer. Here for example, perhaps GM have eventually to license Tesla's self driving software.
BREAKING: General Motors has announced that they have halted funding of its Robotaxi development by Cruise and exiting the Robotaxi business.
— Sawyer Merritt (@SawyerMerritt) December 10, 2024
"GM plans to realign its autonomous driving strategy and prioritize development of advanced driver assistance systems on a path to fully… pic.twitter.com/6180yY8EpG
[11] Google introduces its 'AI agent' that works within the Chrome browser.
Simple example where you give your AI Agent a list of company names in a Google spreadsheet and it goes off and browses the web to get matching contact email addresses. Limitless potential (and dangers). More at Google Blog on Gemini 2.0 Flash including "Unlocking agentic experiences with Gemini 2.0".
Google just unveiled a new project that shows how AI Agents are coming for full browser and computer use. This opens up AI to any of the laborious knowledge tasks that we do today. Most importantly, it lets us automate work that we never had even contemplated before. pic.twitter.com/xLauJhZnr9
— Aaron Levie (@levie) December 11, 2024
[12] Google shows off Project Astra its 'AI assistant' to keep you company
The video shows this AI Assistant being used in the life of an ordinary member of the public.
Here’s a peek at the future: Project Astra, our prototype showing glimmers of a universal AI assistant. We showed an early glimpse of it at I/O, and it’s now in the hands of trusted testers. Here’s how Robbie is using it. Can’t wait til we start shipping, 2025 is going to be an… pic.twitter.com/rci2XeyeHp
— Sundar Pichai (@sundarpichai) December 11, 2024
[13] Concerns about the safety of AI development continue
The competitive market between AI companies means that ever more powerful models are being released every month, most of which now have the power and permisions to control a computer themselves. Here is one safety score just released. If it was a school report then "Could do better ..." might be appropriate.
🆕 Out now: FLI's 2024 AI safety scorecard! 🦺⬇️
— Future of Life Institute (@FLI_org) December 11, 2024
🧑🔬 We convened an independent panel of leading AI experts to evaluate the safety practices of six prominent AI companies: @OpenAI, @AnthropicAI, @AIatMeta, @GoogleDeepMind, @xAI, and Zhipu AI.
🧵 Here's what they found: pic.twitter.com/BCjVAWgPB2
... and this is the rate of increase in the intellectual power AI models
In one year, AIs went from random guessing to expert-level at PhD science questions
— Chubby♨️ (@kimmonismus) December 11, 2024
2025 is going to be an interesting year. pic.twitter.com/lkbdRx9YCQ
[14] AI will at least make us laugh
There will be ever more inventive political movies, at least while social media has X to publish on.
AI is out of control 🤣 pic.twitter.com/DezfS05kec
— Clown World ™ 🤡 (@ClownWorld_) December 11, 2024