Latest ChatGPT updates and Google's response

Hi!

OpenAI completed its festive roll-out this week with five more days of releases and announcements - here's what you need to know about Days 8-12.

ChatGPT Search free to all 

We first talked about the SearchGPT prototype in August and ChatGPT Search was made available to paid users at the end of October. This week brought good news for users on the free plan however as ChatGPT's web search capability is now available to everyone. 

The feature has also received some performance upgrades in response to user feedback with improved speed generally and better performance on mobile.

Search has also been integrated with Advanced Voice mode to give users access to up-to-date web information regardless of how ChatGPT is accessed.

One for the developers

The day 9 announcements may have gone over the heads of those less technically-minded but it was welcome news for developers. In simple terms, OpenAI began the roll-out of some new developer tools, customization options and performance upgrades - including the addition of the 01 model in the API.

They also announced a reduction in costs for some of the GPT-4o options making AI more accessible for more developers. 

Call or WhatsApp ChatGPT

Want to chat with ChatGPT on the go? You can now access it through WhatsApp globally, or via phone call if you're in the US (1-800-242-8478). You don't even need a ChatGPT account to try it out - everyone gets 15 minutes of free calls to test the service. More of a stocking filler than a big announcement but it opens up ChatGPT to a wider audience.

Mac desktop app upgrades

Another niche offering on day 11, this time only really relevant for Mac users. The desktop app helps streamline workflow by reducing the need to switch between multiple tabs and apps and allows you to automate some of your work using ChatGPT. 

The latest update allows the Mac ChatGPT desktop app to work seamlessly with more tools including Apple Notes and Notion. It also incorporates Advanced Voice mode so you can collaborate with ChatGPT in the way that best suits you. Windows users, don't worry - your version is on the way!

A sneak peek of o3

The 12 days of OpenAI began with the full release of the o1 model and we're ending with some previews of o3 and o3 mini (o2 was skipped to avoid confusion with the phone company of the same name).

The release of o3 mini is expected in late January 2025, with the full o3 release following shortly after. However early access is being opened up before then to external safety testers. 

Initial benchmarks show the improvements in coding, math, and science capabilities that we’ve come to expect with each new model. o3 also builds on the enhanced reasoning introduced with o1.

So with day 12 ticked off the calendar, was it all we’d hoped for? There have been big announcements with Sora, Pro and the incoming o3 model but I have to say, the last week has seemed a little underwhelming. 

It’s not to say these aren’t useful additions but it does feel as though the 12 days ended with more of a whimper than a bang.

Competition in the AI space remains fierce

Not to be outdone, Google has been busy with its own December releases, notably with Veo 2 - their answer to Sora in the video generation space. Early user reviews seem to suggest it’s the better of the two which must be a bit of a blow to OpenAI after the long build-up and grand reveal for Sora. 

Two updated Google models were also introduced in the form of Gemini 2.0 Experimental Advanced and Gemini 2.0 Flash Thinking Experimental, with the latter specifically designed to enhance reasoning capabilities and show its thinking process.

Alongside an update for the improved Imagen 3 image-generation model, Google also launched some interesting tools in Google Labs, including Whisk for image blending and an update to the NotebookLM podcast feature. The existing set-up allowed users to input a document and generate a realistic sounding podcast based on the document contents - already an amazing tool. The latest version allows you to interrupt the ‘podcast hosts’ and interact with them in real-time.

So as we head into 2025, one thing's certain - the competition between AI leaders will continue to drive innovation. With companies big and small pushing boundaries in every area of AI, the landscape promises to remain as dynamic as ever.

OK, speak soon!

Inge
Top Apps Founder

P.S. Check out the Top Apps AI courses here.