Sorry for the tardiness of today's episode, a lot happening in the AI space and on my Thursday evenings as well.

THE BIG NEWS STORY TODAY - AU GOVERNMENT AI REGULATION
So the government has announced that after 7 months, its advisory board has come back with some suggested guardrails for AI which will now go through a further round of consultations - meaning no legislation until some time in 2025. In the meantime they have released a voluntary set of AI safety guidelines that businesses can choose to adopt if they want to. This is all against a backdrop of a recent government funded survey of business leaders that found a third of businesses using AI aren't letting customers know they are doing so and have no intention of changing that stance.
SO WHAT: There is movement at the station... but it is very slow. While the EU put their AI laws into place in August we still don't even have a draft of ours, and we won't until some stage next year. To me this mirrors the approach to digital privacy, where Australia has fumbled for more than four years now and fallen behind not only other countries, but more importantly their own citizen's expectations. I haven't had a chance to deep dive into the voluntary code that has been released, but after a quick skim through it all looks very reasonable and makes sense - the problem is without the stick (and no real carrot) the incentive for organisations to make the investments needed to adopt them is very low, and as such they won't have much of a real impact. The AI space is moving quickly, and we need our politicians to put on the afterburners and catch up, it's a complex and difficult challenge to regulate AI, but I worry the horse is already bolting and it might be too late for them to catch up to it.
> Read More
> LINEKDIN: My Rant on This Issue

LOTS OF PEOPLE ARE USING AI TOOLS
Meta announced they have over 185 million active weekly users, and that their open source Llama LLM has been downloaded 350 million times. Meanwhile ChatPGT has hit 200 million active weekly users and also revealed that 92% of Fortune 500 companies are using their AI tools. 
SO WHAT: While you have a bunch of naysayers out there (more on that shortly) questioning whether AI is all hype, it appears a whole heap of people in business are using it regularly. While businesses are still grappling with how to quickly roll out structures, processes, policies and approved tools it seems like people are finding ways to use AI themselves - whether their bosses are aware or not.
> Read More

AUSTRALIAN GOVERNMENT RESEARCH SHOWS AI IS NO GOOD
In news that appears contrary to everything else we read, see or experience, the Australian government through ASIC engaged Amazon to evaluate whether AI is actually any good at summarising information or not. The results, AI is apparently not so good. The "experts" running the tests even went to the trouble of adjusting "behind the scenes" settings such as temperature and top-k (my sarcasm may not come across very well here I know).
SO WHAT: This is another example of how slow and out of touch our government is. They probably paid Amazon hundreds of thousands of dollars to do a pretty basic test and came up with results that are contrary to everyone else's research. The news which has been released this week is about research done at the start of the year using a now extremely outdated AI model and methods not suited to the outcomes they were looking for. Expertise in the AI space is thin on the ground despite the claims various organisations might make.
> Read More

DAWN OF A NEW WAY OF DEVELOPING VIDEO GAMES
Google researchers have created a neural network that can generate real-time gameplay for the classic game Doom - without using a traditional game engine. In another recent study some researchers were able to do a similar thing with Mario Brothers.
SO WHAT: The gaming industry is worth about $400 billion dollars globally... now imagine you can start to create games without having to write millions of lines of code and spend thousands of hours developing complex 3D models. Potentially, in a few years time you might be able to just come up with an idea for a game and then get an AI to make it for you in a few minutes - it's going to be wild.
> Read More

AI SHORTS (LOTS OF THEM)

• Atlassian is going to get into what it thinks is the lucrative AI meeting bot space, but all these platforms are too expensive and there are cheaper and better ways to add this capability (hit me up if you want to know how) - more

•  

  Former OpenAI chief scientist Ilya Sutskever has raised over $1 billion in capital for his 10 person, 3 month old start-up called Safe Superintelligence. Wow. - more
• Anthropic is going enterprise to let Claude compete with ChatGPT for corporate dollars - more
• The EU is trying to fine US start-up Clearview for breaching its AI laws by scraping faces off the web and using them to create a searchable database, Clearview don't care and won't pay - more
• An interesting article that serves as a warning about all these AI platforms that are popping making bold claims with many of them ending up being just a pretty storefront with no tech behind them - more
• Elon Musk has spent circa $3 billion on some new AI hardware to try to get X back into serious contention in the LLM space, and as always he has chosen a name any man-child would be proud of "Colossus" - more
• Looks like Amazon will do a deal with Anthropic to use their Claude LLM in the Alexa voice assistant - more
• A chef in Dubai has admitted a popular new pizza on his menu was the result of a product inspiration session with ChatGPT - more
• While California is about to introduce its own AI regulation, and the US federal government hasn't mentioned what they are going to do, the U.S. Artificial Intelligence Safety Institute announced agreements on AI safety research, testing and evaluation with both Anthropic and OpenAI - more
• Google has rolled out safeguards in its AI products for the US election, which basically ban their AI tools from talking about it in ay way - more
• A start-up in the US is using AI to control what kids see on the internet to make sure it is safe for them... I'm sure this won't end badly - more
• Google is releasing their version of customGPTs, they are called GEMs and they will start to roll out to paid customers soon - more
• We may be seeing the genesis of Robocop, as an oversized robotic witches hat on wheels powered by AI - more

KNOW SOMEONE WHO MIGHT WANT TO SUBSCRIBE?
PLEASE SEND THEM THIS LINK...
> CLICK HERE

That's it for this week. For those asking, yes, all the photos of me in these emails are not really me and are AI generated likenesses. If you have any tips on improvements or content to include then feel free to hit me up - [email protected].