AI update: OpenAI’s ChatGPT-4o, Google Gemini 1.5 Pro and Claude iPhone app

AI update: OpenAI’s ChatGPT-4o, Google Gemini 1.5 Pro and Claude iPhone app

It was a big week in the world of AI. Both Google and OpenAI released major updates. You know I’ve got your back. I cut through their marketing-speak to decipher what these updates actually mean for you and me. 

Gemini AI gets a bigger boat

Google’s AI, Gemini 1.5 Pro, will handle tons of data. Need to summarize 1,500 pages of text? No problem. It’s pretty nuts.

Google goes ‘Ask Jeeves’ mode

You can ask a question in the search bar and it’ll summarize the answer for you. Great for you … not so great for websites that rely on traffic.

  • How to use it: Looking for the best way to clean a pair of leather boots? An AI Overview will give all the detailed steps to get them back in mint condition — without having to click into a single link.

Workspace gets a new intern

Google Workspace’s “AI Teammate” searches your messages, email threads and attachments. Think of it like that coworker who’s been at the company for 20 years and knows every project, process or strategic conversation you might need to reference. 

  • How to use it: Ask AI teammates if you’re on schedule to launch a new product. It’ll pull relevant dates, action items and notes from your Workspace apps and give you a status report in bullet form. Super useful.

Ask Photos, Audio Overviews and Project Astra 

Google also announced a ton of audio/visual-focused AI updates.

  • Ask Photos can help you find an image in your camera roll or answer questions from your photo data. No more scrolling through photos to find that random pic.
  • How to use it: Ask a question like “Show me the best photo from each National Park I’ve visited” or “What was the theme of my nephew’s last birthday party?” Viola!
  • Audio Overviews can create audio discussions based on a text prompt. 
  • How to use it: Want to know the answer to a wacky science problem but can’t read on your commute? Put your question into Google’s chatbot and you’ll get a response through interactive audio.
  • Project Astra: No, it’s not the name of a secret government spy program, Project Astra is Google’s latest iteration of its AI assistant. It can help you find answers via audio and video instead of text, with a lot less lag.
  • How to use it: You can’t … yet. But, in a demo, it helped a user find their glasses, answer questions about speaker parts and even reviewed code. Pretty slick, if it actually works.

GPT-4o gets an empathy upgrade

What has Google pulling out all the stops? OpenAI. In a livestream just before Google’s I/O announcements, OpenAI announced a bigger and better model, GPT-4o

GPT-4o can respond to and generate any combo of voice, image or video — twice as fast as the more recent model. But what got my attention was that it’s also getting more … emotional. It can pick up on different tones in a conversation and respond like it has feelings. 

  • How to use it: You can interrupt GPT-4o in the middle of an answer — and it’ll recalibrate and adjust its response based on your tone. Ya know, how frustrated you are.

You’ve got to watch this video of ChatGPT-4o tutoring a kid through math problems. Wish I had this when Ian was in school … And in this other video, the bot’s lifelike conversation and laugh totally freaked me out.

FYI, I asked, and they told me ChatGPT Plus members will get access to the alpha version in the next few weeks. I pay for it, but I’d say most folks don’t need the upgrade.

Just remember …

No matter how helpful they get, these bots are machines. Always cross-check their advice — and promise me you won’t break into Windsor Castle because a chatbot told you to … Yeah, a British guy really did that last year. 

Tags: Apple iPhone, apps, camera, Google, Livestream, matter, photos, school, updates, upgrades, video