Neural Notes: Ed loves AI, multiple Google geminis and Leonardo.Ai’s $47 million raise

Welcome back to Neural Notes: A weekly column where I condense some of the most interesting and important AI news of the week. And it’s been a big one! Australian AI startup Leonardo.Ai secured a casual $47 million, Ed Husic announced a $17 million ‘AI adopt’ program to aid small businesses, and Google unleashed its long-awaited Gemini AI on the world.

And in other news, Canva released some end-of-year data related to its own AI offerings, and they were admittedly quite impressive. The Aussie unicorn told SmartCompany that its AI products have been used over four billion times. Its overall users have also jumped from 150 million a month back in October to 170 million a month.

You can read the full story here.

Ed Husic is loving an announcement lately, and he’s got another one

The Minister for Industry and Science, Ed Husic, has been on one with end-of-year announcements. Earlier this week it was an $18.5 million grant for the quantum computing sector. And the week before was the $392 million Industry Growth Program which sounds rather similar to the Coalition’s Entrepreneurs’ Programme that was cut earlier this year.

And now as we slide into the weekend, Minister Husic is back with $17 million for the Responsible Artificial Intelligence Adopt Program.

Announced alongside the AIIA, the program will see the establishment of five centres across Australia to assist businesses in safely adopting AI.

According to the AIIA, the centres will provide free-of-charge specialist training to businesses on weaving AI responsibly and efficiently into business operations.

“We want to rebuild our sovereign capabilities here, and also create good jobs in the process. Through using more technology in our businesses, we can strengthen the way they operate, secure jobs and do what we need to do as a country,” Mr Husic said during the announcement in Canberra on Friday.

Minister Husic said the AI sector has grown by roughly 7% over the past year.

“But we know that there are a lot of smaller businesses and medium enterprises that may not be harnessing the full benefit of AI, not using it because they don’t know necessarily how it will work for them, probably have some concerns around whether or not they can manage it themselves, or thinking that it might be a costly investment to business.”

Minister Husic went on to say that more investment into AI is going to be critical for Australian businesses in the future.

“We want strong businesses, we want more work being done onshore, and technology – particularly AI – that can crunch through large lots of data very quickly, very efficiently, and be able to give you insights that would take way longer than what people would be able to do on their own.”

The Labor government is currently looking for applications from Aussie businesses that specifically target one of the priority areas in the National Reconstruction Fund (NRF):

• Renewables and low emissions technologies
• Medical science
• Transport
• Defence

“It is estimated that more than a quarter of the Australian economy will be rapidly and significantly disrupted by Generative artificial intelligence, which means nearly $600 billion of economic activity faces Gen AI disruption. This AI Adopt Program makes sure Australia is ready to ride the productivity wave,” AIIA CEO, Simon Bush, said.

“Australia ranks near the bottom of business AI adoption tables globally, so this new government program being announced today is a necessary and important investment by the government to ensure Australian businesses have the skills and confidence to adopt generative AI technologies.”

The AIIA has also announced a partnership with the Queensland government to deliver grants for AI ‘microcredential courses’ for SMEs.

According to the AIIA, the program will allow for 650 small Australian businesses to enroll in a 12-week, self-paced course on the adoption of generative AI.

Hey Leonardo (She LAikes me for me)

Sticking with Australia, the biggest news of the week has been the mammoth $47 million raise for fledgling generative AI startup, Leonardo.Ai.

Based in Sydney, Leonardo.Ai initially targeted the video game industry with its hyperrealistic image generation technology. But it very quickly pivoted into other creative sectors such as marketing, fashion, and advertising.

And it’s worked, nailing two successful raises in 2023, with Side Stage Ventures confirming it participated across its Seed and Series A rounds this year. According to the VC, it was convinced by the tech back in February — just six weeks after its launch.

The startup’s AI leverages open-source technologies and is trained with synthetic and Creative Commons data, offering features like editing and creation of consistent style assets, along with a user-friendly API.

Since launching in 2022 it has produced over 700 million images and currently generates about 4.5 million images daily. That’s huge, so it’s no surprise that it has also expanded its offerings to include an enterprise tier. It is firmly targeting the SME sector, with collaboration tools and private cloud hosting. There are also plans to develop more B2B features so stay tuned for that one.

The sheer volume of users also resulted in Leonardo publicly partnering with AWS to handle the influx and remove the cap it had on its waiting list.

The company also briefly mentioned plans to release a Shield product that will provide indemnification for commercial clients. We are yet to get any further details on when this might drop or how much it will be, but we’re keeping an eye on it.

There’s a lot more to the Leonardo.Ai announcement, and we have the full story here for you. Definitely an AI startup to watch in 2024.

And since we’re speaking about generative AI photos, you should know that the term ‘Promptography’ now exists in the world thanks to an artist in Germany. The term seems to be applied to someone who pops text prompts into image-based generative AI platforms and further tweaks them to create “thought-provoking photographs”.

As someone who physically shudders over hamfisted attempts to mash business words together (think ‘mumpreneur’ and ‘regtech’), this is upsetting to me on a spiritual level.

It’s (Google) Gemini season

Internationally, the big news of the past 24 hours has been the release of Google Gemini — which promises to significantly enhance the functionality of the tech giant’s chatbot, Bard.

This integration is expected to modify Bard’s capabilities in areas such as reasoning, planning, and understanding.

The long-awaited AI, as it turns out, is actually a suite of AI that sounds like a range of iPhone products. The idea is for Gemini to be scalable — usable across large data centres as well as on personal mobile devices.

• Gemini Ultra — the big boy flagship
• Gemini Pro — the mid-range model
• Gemini Nano – designed for mobile devices, with the Pixel 8 Pro already confirmed.

And if that wasn’t already confusing enough, the Nano has two sizes to accommodate higher and lower memory devices. There are also plans to roll out Bard Advanced in 2024, which will utilise an enhanced version of Gemini.

Google has claimed that internal benchmarks have revealed Gemini to outpace OpenAI’s GPT-3.5 across several tests. However, there have been mixed reactions from users in the wild so far.

According to tweets sources by Tech Crunch, some Gemini Pro users have been reporting it getting quite basic information wrong.

I'm extremely disappointed with Gemini Pro on Bard. It still give very, very bad results to questions that shouldn't be hard anymore with RAG.

A simple question like this with a simple answer like this, and it still got it WRONG. pic.twitter.com/5GowXtscRU

— Vitor de Lucca 🏳️‍🌈 (@vitor_dlucca) December 7, 2023

Considering that the original release of Bard saw actual Google employees call it “worse than useless” and a “pathological liar” — the latest problems have gotten a firm yikes from me. Here’s to hoping these bugs get worked out before the wider release of the various Geminis.

For now, the new Bard is available in English in 170 countries, with plans for further language and regional expansions. There are plans to include multimodal input capabilities (like what ChatGPT now allows) and undergo additional safety evaluations for Bard Advanced.