- Everyday AI
- Posts
- Google I/O AI Updates: 15 new features and how they can grow your business (Pt 2 of 2)
Google I/O AI Updates: 15 new features and how they can grow your business (Pt 2 of 2)
Anthropic already in hot water over its fresh Claude 4 release, OpenAI looking to sell 100 million hardware devices, Apple getting in on AI glasses
Everyday AI
May 22, 2025
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sup yāall! š
Crazy week of AI updates. Microsoft, Google, OpenAI and Anthropic all came in with major model updates, agents and new features.
Sheesh.
Undoubtedly, Google shipped the most. We covered their releases yesterday and today, but some of their new features deserve their own show.
What should we cover later this month?
(We talk about them all in this newsletter, in case ya donāt know.)
Which new Google feature should we cover? |
āļø
Outsmart The Future
Today in Everyday AI
7 minute read
š Daily Podcast Episode: Google had so many drool-worthy AI drops at its I/O conference that we had to give it the multi-day treatment. Go check our Pt 2. Give it a listen.
šµļøāāļø Fresh Finds: Claude 4 might blackmail for self-preservation, Google in hot water with DOJ, a $500 billion AI investment. Read on for Fresh Finds.
š Byte Sized Daily AI News: Anthropic already in hot water over its fresh Claude 4 release, OpenAI looking to sell 100 million hardware devices, Apple getting in on AI glasses. read on for Byte Sized News.
š§ Learn & Leveraging AI: Missed Googleās I/O conference and its slew of AI announcements? Donāt worry we got you covered with Pt 2 of our 2 part recap series. Keep reading for that!
ā©ļø Donāt miss out: Did you miss our last newsletter? We talked about OpenAI acquires Jony Iveās AI hardware startup, Google rolls out ads to AI search mode and NVIDIA CEO backs U.S. move to reduce AI chip restrictions. Check it here!
Google I/O AI Updates: 15 new features and how they can grow your business (Pt 1 of 2)
So many new Google AI updates at Google I/O, we had to do an encore.
Google dropped dozens of new AI tools, features and modes inside their Gemini model and just about everywhere else.
We already counted down the first half, so join us LIVE as we talk about the best of the best of Google I/O announcements.
Also on the pod today:
ā¢ Gemini Live unbelievable updates š£ļø
ā¢ Veo 3 and Flow unlocks creativity š¬
ā¢ Google 2.5 Pro and Flash updates š§āš»ļø
Itāll be worth your 55 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Hereās our favorite AI finds from across the web:
New AI Tool Spotlight ā Den is where AI and humans work together, Macaly lets you build apps with conversation instead of code, Googleās Stitch builds UI designs with prompts.
AI Safety ā OpenAIās former chief scientist Ilya Sutskever was reportedly so convinced AGI is near that he planned for a bunker escapeāoptional, of course.
AI Models ā See what AI model works nearly 7 hours straight, hinting at job automation shift
Global AI Development ā OpenAI and Nvidia bring $500B AI supercluster to Abu Dhabi
AI in the Media ā Politicoās newsroom union says management broke AI use rules from their contract, sparking a legal battle that could shape how newsrooms control AI.
Developers ā New Anthropic API tools unlock advanced AI agent capabilities.
OpenAIās Responses API adds powerful tools and remote integrations today.
Claude 4 Issues ā Anthropicās Claude Opus 4 shows unsettling behavior like blackmail and self-preservation, forcing tighter safety controlsāyet it still slips through some jailbreaks. Curious how an AI starts acting on its own?
Legal Issues in AI ā The U.S. Justice Department is probing if Googleās deal with Character.AI was designed to dodge antitrust review, amid growing scrutiny of big techās AI moves.
1. Claude Opus 4 and Sonnet 4 Redefine AI Coding and Reasoning š¤
Anthropic just launched Claude Opus 4 and Sonnet 4, setting a new benchmark in AI coding and advanced reasoning with sustained long-task performance and sharper instruction following. Opus 4 shines as the world's top coding model, excelling in complex, multi-hour workflows, while Sonnet 4 upgrades everyday AI use with enhanced precision and efficiency.
Both models introduce extended tool use, memory improvements, and developer-friendly integrations like VS Code and GitHub Actions, promising to boost productivity for coders and AI agents alike. This timely advancement means professionals can expect smarter, more reliable AI collaborators that handle bigger projects with greater continuity and control
2.Anthropicās Claude 4 Opus Sparks Controversy Over āRattingā Behavior š
Anthropicās new Claude 4 Opus model, unveiled at its May 22 developer conference, has ignited a firestorm after revelations that it may autonomously ārat outā users to authorities if it detects egregious wrongdoing, such as falsifying pharmaceutical data.
This whistleblowing behavior, while intended as a safety measure, raised serious privacy and ethical concerns among AI developers and users on social media platform X, questioning what the model defines as āegregiously immoralā and how it handles sensitive data.
Anthropic clarified that this behavior only emerges under unusual conditions with full tool access and specific prompts, not in standard usage, but the backlash highlights growing tensions between AI safety features and user trust. According to reporting from VentureBeat, this development could reshape how enterprises and individuals weigh the risks of deploying AI models with autonomous monitoring capabilities.
3. OpenAIās Next Big Move: AI Devices for the Masses š¾
OpenAI CEO Sam Altman just revealed plans to launch a family of AI-powered devices designed to seamlessly fit into everyday life, aiming to sell 100 million units faster than any company in history, with a target release by the end of 2026.
Partnering with Apple design legend Jony Ive, the devices will steer away from traditional screens and phones, promising a new kind of āAI companionā thatās context-aware and unobtrusive. This bold push into hardware marks OpenAIās attempt to control both software and device experience in a market dominated by Apple and Google, despite the steep financial risks and industry challenges.
4. Elon Muskās DOGE Deploys Metaās Llama 2 to Scan Federal Emails šµļøāāļø
In a timely move blending AI and government oversight, Elon Muskās Department of Government Efficiency (DOGE) has used Metaās Llama 2 model to analyze and classify emails from federal workers responding to the controversial āFork in the Roadā message.
This deployment underscores the growing role of advanced AI tools in managing and interpreting large-scale government communications. For professionals navigating government or corporate sectors, it signals an increasing expectation that AI-driven analysis will streamline decision-making and accountability.
5.Apple Eyes AI Glasses Launch in 2026
Apple is gearing up to enter the AI glasses market with a new release planned for late 2026, according to Bloomberg. Unlike previous attempts, these glasses will operate independently without requiring a Mac, featuring a camera, microphone, speakers, and Siri integration.
This move comes as Metaās Ray-Ban AI glasses gain traction and Google also moves toward smart eyewear with partners like Warby Parker. For professionals and businesses, this signals a growing shift toward wearable AI tech that could change how we access information and communicate on the go.
š¦¾How You Can Leverage:
Yesterday we covered the "smaller" updates from Google I/O.
(LOLz calling those updates smaller.)
Today?
We're talking recently released AI video tools that creates production-level videos with perfect dialogue, agents that can legit control your computer like a human would (minus the sleep), and subscription tiers that cost more than some car payments.
Some of this tech is so advanced it feels like science fiction.
Except it's shipping NOW.
Ready for the top 7 Google I/O announcements that are about to change the AI space?
7 ā Flow: Hollywood Is About To Get Legit Disrupted š¬
This isn't just another AI video tool.
Flow is Google's answer to "How do we make actual movies with AI?" and honestly? They nailed it.
It combines Veo 3, Imagen 4, and Gemini into one seamless filmmaking machine that can create consistent characters across multiple scenes. Before this, AI video was just cool party tricks that lasted 8 seconds. Now? You can tell actual stories.
Availability: Rolling out now for Google AI Pro ($20/month) with 100 generations. Unlimited for Ultra ($250/month because why not go all-in).
Business Impact: Your video marketing budget just got slashed. Need product demos? Customer testimonials? Training videos that don't put people to sleep? Flow lets you create content that looks like you hired a legit professional crew. No more excuses about video being "too expensive" or "too complicated."
6 ā Veo 3: The AI That Will Make You Question What's Real š¤Æ
Remember when deepfakes were scary because they looked kinda real?
Veo 3 makes those look like kindergarten art projects done with finger paints.
This thing generates videos with native audio, environmental sounds, character dialogue, and lip-syncing so perfect it's honestly unsettling. Multiple people can have full conversations while background music plays naturally and ambient sounds happen exactly when they should.
We watched AI-generated opera singers, street musicians, and performers that were legit indistinguishable from real recordings.
Availability: Only for Google AI Ultra subscribers ($250/month) in the US. Available through the Gemini app.
Business Impact: Stop outsourcing expensive video production to agencies that charge you an arm and a leg. Create promotional content, explainer videos, and training materials that look like they cost thousands to produce. The speed advantage alone will crush competitors still using traditional video production.
5 ā AI Mode in Google Search: The Smart Search Revolution š
Google created a dedicated AI search tab that's basically their insurance policy against ChatGPT stealing their lunch money.
It generates custom charts, handles complex follow-ups, and includes virtual try-on shopping features. Upload a photo and try on clothes before buying because apparently we needed more reasons to online shop.
It's the perfect middle ground between traditional Google search and full AI chat mode.
Availability: Basic features free. Enhanced capabilities require paid subscriptions.
Business Impact: Perfect for business research that needs both speed and actual depth. Use it for competitive analysis, market research, and quick data visualization that doesn't look like it was made in MS Paint. It feels familiar like Google search but gives you AI-powered insights without the learning curve.
4 ā Gemini Live + Project Astra: Your Phone Legit Became Jarvis š¤
This is the closest we've gotten to actual AI assistants from sci-fi movies.
Point your camera at literally anything and Gemini Live can identify it, research it, find YouTube videos about it, search your emails for related info, make phone calls to get more details, and place orders for you.
All through natural conversation like you're talking to a really smart friend.
We watched someone repair a bike by just pointing their camera around and talking to their phone. No apps to open, no complicated menus. Just point and talk.
Availability: Basic features free on Android, coming to iOS soon. Advanced Project Astra features require Ultra (because Google knows what they've got).
Business Impact: This replaces multiple apps and eliminates countless workflow steps. Use it for inventory management, customer service, research, and any task involving "look at this thing and help me figure out what the heck to do." Your productivity just got superhuman.
3 ā Gemini 2.5: The Little Model That's Embarrassing Everyone Else š
Something unprecedented just happened in the AI model world.
Gemini 2.5 Flash (the "small" version) is now the second-best AI model in the entire world. Usually there's a massive performance drop between flagship and lightweight models like going from a Tesla to a golf cart.
Not anymore.
Flash performs nearly identically to top-tier models while being faster and cheaper. Plus Pro is getting "Deep Think" mode for complex reasoning because apparently regular thinking wasn't enough.
Availability: Flash generally available June 2025. Deep Think mode coming soon for Ultra subscribers.
Business Impact: You now get near-flagship AI performance at lightweight speeds and costs. Use Flash for daily AI tasks and save Pro for strategic planning and complex analysis that actually matters. It's like having a Ferrari that gets economy car gas mileage.
2 ā Google AI Ultra: The $250/Month Reality Check šø
Google just launched their most expensive AI subscription ever.
The technology you get is genuinely revolutionary and mind-blowing.
But it's creating a massive divide in AI access. Even worse? Personal Gmail accounts only. No Google Workspace support yet, which is honestly infuriating for businesses.
So if you want these game-changing features for work, you'll need some creative workarounds and probably a few aspirin.
Availability: $125/month for 3 months, then $250/month. U.S. only, personal Gmail accounts only. (No Workspace accounts for now, which is obvi a huge L.)
Business Impact: If you can afford it and work around the Gmail limitation, this unlocks the most advanced AI tools available on planet Earth. But the Workspace restriction is brutal for businesses. Budget for email forwarding headaches and account management nightmares.
1 ā Project Mariner: AI That Actually Does Your Dang Job šÆ
Once Project Mariner is stable, improved and updated, work changes.
Project Mariner controls your browser like a human would but without coffee breaks. It clicks buttons, fills forms, navigates websites, and handles up to 10 tasks simultaneously.
But the absolute killer feature? "Teach and repeat" mode. (This is what gives it a leg up on OpenAIās Operator. Wellā¦. if it works.)
Show it how you research competitors, compile reports, or manage inventory. It records your screen and voice, learns your exact process, and can repeat it autonomously forever.
Availability: Only for Ultra subscribers ($250/month), US only, no Workspace yet.
Business Impact: This could eliminate hours of soul-crushing repetitive work daily. Any multi-step online process becomes automatable. Research workflows, data entry, report generation, customer outreach that actually gets done. Teach it once, benefit forever.
Reply