AI’s Great Battle for your Web Browser
Tonnes of releases from OpenAI, Google, Microsoft, and Meta all highlight that 2024's battle will be to become your primary window for everything
Whether it was coordinated or an accident, December has been flooded with new AI product releases. The most notable has been OpenAI which has gone for the wow factor with their 12 days of releases. Google followed this with big headlines about Gemini 2.0 and entering its ‘agentic era’ which feels more like a retaliatory strike against recent ChatGPT and Claude buzz.
What does this all mean to us, the product, design, and strategy leaders deciding how to work with AI?
Have any questions about AI you wanted answered on Design of AI podcast?
Email us info@designof.ai or message me on LinkedIn. We’ll cover questions in the year-end episode.
1. The cost and barrier to entry to use AI continue to drop
AI makers are burning way cash at a pace never seen before in history. OpenAI, Anthropic, and xAI have raised upwards of $20B in the last few months alone. They’re desperate to grow their user base so they can show more revenue on their balance sheet.
Meanwhile Meta and Google are undermining everyone by making AI cheaper and easier to access. This is a win for AI consumers but also a reminder that the financials of AI will be a bloodbath for quite a while as these corps battle to the utilities provider you rely on most for your next product.
2. UX is defining the next generation of AI products
OpenAI’s Canvas update is a game changer in a lot of ways:
Enables collaboration
Empowers custom GPTs
Competes with coding assistants
Turns ChatGPT into a platform
This is great news for the designers who have been complaining that AI products have been strung together without consideration for the user. The evolution happening within ChatGPT and with Claude’s Artefacts has been especially astounding.
Slowly AI products are moving away from chat-based interfaces. Remember when you had to use Discord to use Midjourney?
Notable interfaces now include NotebookLLM, Cove, Make, Superads, Aftercare and more coming out daily.
The downside is that because of the generative capabilities of these apps, you need fewer designers to work on these products than you normally would.
3. AI models are getting better the models aren’t valuable enough to be mass consumed
OpenAI’s new video generation platform Sora is very impressive.
But is it useful? This is the question plaguing businesses needing to monetizing AI. Another example is Suno, the music generation platform. I love playing with it to make silly songs for my daughters and to make my friends laugh. But its not useful, yet.
So as much as AI models look mature and ready to change the world, the great leap towards AI being valuable will depend on developers creating niche applications by appending to the model.
An example of what success would look like is Archetype AI which collects real world data from sensors and environments and leverages the models to evaluate and predict scenarios.
Real value requires leveraging your first party date to bring unique context. And orgs are quickly going to recognize the importance on the semantics of making sure the first party data is easily explainable.
4. This is an arms race to become your primary window for everything
The US Justice Department wants to break up Google because of they dominate the ads, search, and browser spaces. Effectively, no matter where you are in the world, or what type of device you’re on, you’re likely using Google for least one layer of the conversion funnel.
But that dominance may come to an end. Perplexity, OpenAI, Microsoft are directly challenging on search. And now the ‘agentic era’ will challenge Google’s dominance on browsers.
AI agents have the ability to execute tasks for us and on our behalf. They would have the potential to change UX in some significant ways:
Act as the oracle for any questions about your data, apps, requests
Remove the need to go to a browser to seek solutions
Remove the need for us to do the work, just review the answers
All of these are terrible news for Google and why they’re throwing their hat into the Agent ring alongside Salesforce, Microsoft, and everyone else. Here’s a demo of their Project Mariner Chrome plugin that operates within the browser.
Conclusion
This has been an absolutely huge month for AI model makers. Everything related to AI is getting easier. Ultimately this is great for businesses and developers that want to leverage the tech and seek out opportunities to deliver value. But we’re also about to see major disruptions to human-computer interactions that we’ve become very accustomed to.
There’s an expectation that agents will replace a lot of the mundane work we do. But if we take a look at creative-generation apps like Suno and Sora, the question becomes whether anyone will pay for creativity at scale once most of it has become automated? And with AI tokens continuing to drop in price its much more cost-effective for orgs to outsource decisions and work to the tech.