The IT Nerd

Today, Google made a series of announcements about a major AI breakthrough which is their largest and most capable AI model, called Gemini. 

The news spanned the world of enterprise, developers and consumers, so I thought it would be helpful to summarize the main announcements and provide links to the most useful blog posts. 

Today’s news 

Today, Google announced Gemini — the most capable general AI model they have ever built. It is the result of large-scale collaborative efforts by teams across Google, including Google DeepMind and Google Research, and is their largest science and engineering project ever. 

Google has optimized Gemini 1.0, our first version of the model, for three different sizes:

• Gemini Ultra — their most capable and largest model for highly-complex tasks
• Gemini Pro — their best model for scaling across a wide range of tasks
• Gemini Nano — their most efficient model for on-device tasks

What is Gemini? 

Gemini is a multimodal AI model. This means that it can generalize and seamlessly understand, operate across and combine different types of information, including:

• Text
• Images 
• Audio 
• Video 
• Coding languages

It’s also their most flexible model yet, able to efficiently run on everything from mobile devices to data centres. Gemini will significantly enhance the way developers and enterprise customers build and scale with AI. 

Built on next-generation capabilities

Until now, the standard approach to creating multimodal models involved training separate components for different modalities and then stitching them together to roughly mimic some of this functionality. These models can sometimes be good at performing certain tasks like describing images, but struggle with more conceptual and complex reasoning. 

So Google designed Gemini to be natively multimodal — pre-trained from the start on different modalities. Then they fine-tuned it with additional multimodal data to further refine its effectiveness. This helps Gemini seamlessly understand and reason about all kinds of inputs from the ground up, far better than existing multimodal models — and its capabilities are state-of-the-art in nearly every domain. 

Learn more about Gemini’s capabilities and see how it works.

Benchmarking tests

Google has been rigorously testing our Gemini models and evaluating their performance on a wide variety of tasks. From natural image, audio and video understanding to mathematical reasoning, Gemini Ultra’s performance exceeds current state-of-the-art results on 30 of the 32 academic benchmarks widely used in large language model research and development. 

You can see more details in this technical whitepaper. 

Making Gemini available to the world

Gemini 1.0 is now rolling out across a range of products and platforms:

For consumers

• Starting today, Bard — using  a fine-tuned version of Gemini Pro — will be available in English in more than 170 countries and territories. It will be far more capable at things like understanding and summarizing, reasoning, brainstorming, writing and planning. Google is enthusiastic about bringing Bard’s generative AI potential to Canadians soon; 
• Google is also bringing Gemini to Pixel. Pixel 8 Pro is the first smartphone engineered to run Gemini Nano, which is powering new features like Summarize in the Recorder app, and rolling out in Smart Reply in Gboard, starting with WhatsApp, with more messaging apps coming next year; 
• And in the coming months, Gemini will be available in more of their core products and services like Search, Ads, Chrome, and Duet AI.

For developers

• Starting on December 13, developers and enterprise customers can access Gemini Pro via the Gemini API in Google AI Studio and Vertex AI:
• Google AI Studio is a free, web-based developer tool that helps developers and enterprise customers prototype and launch apps quickly with an API key; 
• When it’s time for a fully-managed AI platform, Vertex AI allows customization of Gemini with full data control and benefits from additional Google Cloud features for enterprise security, safety, privacy, and data governance and compliance.  
• Android developers will also be able to build with Gemini Nano, their most efficient model for on-device tasks, via AICore. AICore is a new system capability available in Android 14, starting on Pixel 8 Pro devices. Sign up for an early preview. 
• And as part of their extensive trust and safety checks for Gemini Ultra, they will make it available to select customers, developers and partners for early experimentation and feedback before making it broadly available to developers and enterprise customers early next year 

Looking ahead 

This is a significant milestone in the development of AI, and the start of a new era for Google as they continue to rapidly innovate and responsibly advance the capabilities of our models. They’ve made great progress on Gemini so far and they’re working hard to further extend its capabilities for future versions. 

On the one year anniversary of ChatGPT going public comes these recent findings by Google researchers on ChatGPT’s training data.

The researchers successfully prompted ChatGPT to disclose parts of its training data using a novel attack technique, which involved asking the chatbot’s production model to repeatedly echo specific words indefinitely.

Anurag Gurtu, CPO, StrikeReady had this to say:

The exposure of training data in ChatGPT and other generative AI platforms raises significant privacy and security concerns. This situation underscores the need for more stringent data handling and processing protocols in AI development, especially regarding the use of sensitive and personal information. It also highlights the importance of transparency in AI development and the potential risks associated with the use of large-scale data. Addressing these challenges is critical for maintaining user trust and ensuring the responsible use of AI technologies.

This is not a good look for AI in general and ChatGPT specifically. Clearly people behind AI products need to get a handle on this sort of thing quickly or these sorts of issues will simply multiply.

UPDATE: Kevin Surace, Chair, Token adds this:

The attack was incredibly simple and some of them still work as of now. It is an absolute disaster for any model to reveal its training data – IP-wise, legal, integrity and so on. Certainly, OpenAI and others must put in more stringent safeguards to keep this from happening again.

If you are a user of the Chrome browser, you should update it ASAP to stop a flaw in said browser from being used by threat actors as an attack vector to pwn you.

The details of the flaw can be found here. The flaw allows an attacker to execute code and pwn you. Which is of course bad. What’s worse is that this exploit according to 9to5Mac is actively being exploited. Making this a today problem for you.

Time to update all the things Chrome related.

Shopping habits have significantly shifted over the years as Canadians become increasingly digitally savvy shoppers. With Black Friday and Cyber Monday around the corner, Google Canada is revealing new insights around how Canadians plan to shop this holiday season, and the top searched products under a number of categories. 

Whether Canadians are looking to revive cargo pants, up their hair and nails game or contemplate whether the iconic Mattel hunk is truly “Kenough” for gifting this holiday season, Google Canada’s has the top searched items for holiday shoppers along with additional insights on how Canadians are shopping this season here.

Here’s what Canadians are searching for leading into Black Friday/Cyber Monday*:

Top trending toy searches Dune Lego Blox Fruits Plush Pomni Plush OVO BearbrickToddler Balance Bike Wood Play Kitchen Marvel Action Figures BarbieTrain Set Ken	Top trending apparel searches Crocs Carhartt beanie Vessi shoes Birkenstock Clogs Cargo pants Ugg slippers Nike Air Force 1 Patagonia Fleece Lululemon Bag Loafers
Top trending consumer electronic searches Dyson headphones Bose soundbar Remote car starter Xbox 360 iPhone 15 Pro Max Nintendo Switch Games Garmin Forerunner 45 Watch Sonos speakers Sony Alpha Camera Kindle	Top trending beauty searches ELF lip oil Dior lipstick Clinique black honey Nail stickers Bubble moisturizer Press on nails Aquaphor lip balm EOS body lotion La Roche-Posay Cicaplast Baume B5 Maybelline vinyl ink

Source: Google Search Trends.*

Today, Google Canada is announcing the winning Canadian advertisers and agencies selected for the inaugural Google Search Honours Awards, celebrating those who are using AI-powered advertising to propel their business forward. 

There are plenty of awards that recognize flashy brand campaigns, but Search marketing doesn’t get the same recognition despite its impact on driving business results. The Google Search Honours Awards were created to celebrate brands and marketers using Google’s AI-powered ads to reach new audiences, maximize their spends and push the boundaries of what their campaigns can achieve.

On September 14, Google announced the shortlist, and six winners were announced at the awards celebration on Thursday November 2.

Category	Award	Candidates
Reach Expansion The Reach Expansion category recognizes work that leverages AI-powered Search tools to efficiently connect with a wider audience.	Rising Star	Winner: Simplii Financial Shortlist: BMO | KINESSO; Desjardins | Glassroom; Hyundai | INNOCEAN
	Best in Class	Winner: Canadian Tire Shortlist: Hudson’s Bay; lululemon athletica; RBC Insurance | WebAgency
Value Impact The Value Impact category celebrates AI-powered work that uses smart bidding strategies to get the most out of Search advertising investments.	Rising Star	Winner: Shopify Shortlist: Destination BC | Noise Digital; Scotiabank; WestJet Airlines Ltd. | Touché!
	Best in Class	Winner: Toyota | Saatchi & Saatchi Shortlist: Air Canada; Jewlr; QuickBooks Online
AI Innovation The final category is AI Innovation, which shines a light on those who are pushing the boundaries and breaking new ground with AI-powered Search campaigns.	Agency of the Year	Winner: Mindshare Shortlist: PHD; Saatchi & Saatchi/Synergize; Starcom
	Marketer of the Year	Winner: QuickBooks Online Shortlist: Canadian Tire; lululemon athletica; TD Bank

The news is out from Google this morning Demand Gen will enter global general availability for all advertisers.

YouTube & Google announced the beta for Demand Gen back at Cannes – unveiling it as a new ads campaign type for social marketers. Since then, global brands and agencies like Samsung Germany and pHd Media have used this solution to uplevel their approach to social.

Their findings: Neither Google nor YouTube are markedly social platforms but the interconnected networks provide performance and results that are better for social buyers. It’s because of the unparalleled reach (read: billions of users a month) across multiple surfaces and screens, plus thecontent that viewers seek out, and the singular nature of how this ecosystem is where consumers turn to first to engage during critical moments of consideration – from brand letters in Gmail to creator recommendations on YouTube and Shorts.

More details can be found in this blog post.

Today, an Economic Impact Report was released, revealing that Google Search, Google Play, YouTube, Google Cloud and Google Advertising tools delivered over $45 billion of economic activity for Canadian businesses, non-profits, publishers, creators and developers in 2022.

The report, published by Public First, also offers a future-looking perspective on how AI-powered technologies are set to provide the country with an economic boost, whether that’s helping people to work more efficiently, making careers more accessible or unlocking new ways of learning digital skills. Some of the report’s key findings include:

• Generative AI has the potential to increase Canada’s economy by $210 billion –  that’s the equivalent of 8% of Gross Value Added (GVA)
• Generative AI could save the average Canadian worker over 100 hours a year
• It’s estimated that Google Cloud is saving 120 million business hours for the Canadian economy each year – that’s equivalent to at least $7.5 billion a year in business time savings for the Canadian economy.

 Google’s blog post has more information about today’s announcement and you can read the full report here. 

Today, YouTube ads unveiled Google-AI powered solutions that will allow advertisers to effectively unlock creativity on their platform.

The new creative guidance tool in Google ads will auto detect if uploaded campaigns are up to the standard of YouTube’s creative best practices – providing them with input on what is most effective on the platform. This means advertisers will have data-backed input on their campaign impact on the platform – straight from the source. 

This feedback is then made immediately actionable with advertisers ability to then access more AI-powered tools including the ability to add voiceover, perfectly edit their ads for length or orientation, and/or create new immersive and effective ads by adding their creative assets to new templates. 

Once advertisers have the ideal ad, they can trust our campaign types to then place their campaigns, through AI, in the most relevant context for the multi-format, multi-screen experience that is YouTube.

Read more on the blog here.