- Everyday AI
- Posts
- EP 619: Nano Banana Uncovered: A practical guide from inside Google
EP 619: Nano Banana Uncovered: A practical guide from inside Google
Gemini tips from Google Insider, Accenture axing workers who donāt upskill with AI, Meta needs Geminiās help, OpenAIās new Pulse product and more.
Everyday AI
September 26, 2025
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sup yāall! š
Need a double dose of AI for the weekend?
Todayās show with Googleās Paige Bailey was a legit banger. She built a Nano Banana powered app live in literally two minutes so you can watch and follow along.
I was lucky enough to be featured on the WWT AI Proving Ground podcast recently. If youāve got extra time this weekend, give it a listen as we tackled the future of AI in the Enterprise.
See ya Monday. TONs of AI news to tackle. Yikes.
āļø
Jordan
Outsmart The Future
Today in Everyday AI
7 minute read
š Daily Podcast Episode: Ever wish you had a Google leader sit down and explain AI to you with mind-blowing demos? Oh, thatās what we did today. Give it a watch/read/listen.
šµļøāāļø Fresh Finds: Sam Altman on AI vs human intelligence, ChatGPT updates Projects feature, the cost of AI workslop and more. Read on for Fresh Finds.
š Byte Sized Daily AI News: Accenture axing workers who donāt upskill with AI, Meta needs Geminiās help, OpenAIās new Pulse product and more. Read on for Byte Sized News.
š§ AI News That Matters: Think you know Gemini and Nano Banana? Think again. We went hands-on with some impressive use-cases and we unlock it all for you. Keep reading for that!
ā©ļø Donāt miss out: Did you miss our last newsletter? We talked about RSL fights AI scraping, OpenAIās $6.5B and $100M deals, Microsoft Copilot gets official with Claude, Elon goes after Microsoft with AI and more! Check it here!
Nano Banana Uncovered: A practical guide from inside Google š
Everyone knows Google's Nano Banana is bonkers good. š
But did you know you can create an app in minutes that embeds Nano Banana.... and it takes zero coding experience?! š¤Æ
If you haven't used Google's Gemini 2.5 Flash (AKA Nano Banana), you're in for a treat as Google's Paige Bailey gives us the insider's guide.
Also on the pod today:
ā¢ Build apps with plain English š¬
ā¢ Gemini powers Google Sheets magic š„
ā¢ No-code hackathons: non-techies win š
Itāll be worth your 29 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 Scrumball is an AI Influencer Platform for Marketing, Fakeradar helps you stop DeepFake video calls, Neutron is a desktop AI app that proactively helps you before you know you need it.
ChatGPT updates ā ChatGPT rolled out the ability to share projects and (more importantly) said that theyād make connectors available soon in projects and GPTs.
Training AI Models ā Itās getting harder for AI companies to evaluate models because theyāre getting better at humans at real world knowledge work.
AI and the environment ā Amazon is trying to combat the water footprint of using AI. Hereās how.
AI Intelligence ā Will AI surpass human intelligence in 5 years? Sam Altman thinks so.
AI Evaluations ā OpenAI introduced GDPval, a new benchmark of 1,320 real-world, economically valuable tasks across 44 occupations. Find out where your favorite model landed.
AI Slop ā Apparently businesses arenāt seeing ROI on AI because of Workslop, this new study finds.
AI Development ā OpenAI had a monumental week in terms of partnerships. Is it crazy?
1. Accentureās AI pivot speeds up, jobs on the line ā”
Accenture is moving quickly to an AI-first model and plans to āexitā employees who cannot reskill with AI while pouring savings back into growth and talent. CEO Julie Sweet said 550,000 staff have been trained on generative AI, the company launched a six-month $865 million optimization program tied to severance and reductions, and it expects over $1 billion in savings to reinvest.
Accenture now has 77,000 AI and data pros and still plans to grow headcount in the U.S. and Europe after reporting $69.7 billion in revenue, up 7% year over year. Translation for careers and companies: AI fluency is becoming a minimum requirement, and the window to reskill is closing fast.
2.Meta eyes Googleās Gemini use to boost ad targeting šÆ
According to The Information, Meta is in early talks with Google Cloud to potentially fine-tune Gemini and Gemma on Metaās ad data, aiming to boost ad targeting across Facebook, Instagram and WhatsApp. The discussions, which may not lead to a deal, hint at Metaās scaling challenges despite heavy investment, and show the company is open to rival tech if it delivers measurable gains.
If it happens, the move could reshape how ads are targeted across major platforms and signal a more pragmatic era of cross-competitor AI partnerships.
3. OpenAI ships ChatGPT Pulse in preview to Pro users ļæ½*
OpenAI has launched ChatGPT Pulse in preview on mobile for Pro subscribers, a feature that quietly does overnight research using your chat history, memory, feedback, and optional app connections to deliver personalized updates the next day.
Pulse can prep you for meetings by connecting to Google Calendar, draft agendas, suggest dinner ideas, set reminders, and surface next steps on longer-term goals, with all integrations off by default until you opt in. OpenAI says users can steer what Pulse tracks so updates stay focused, a bid to make ChatGPT feel less like a static assistant and more like a daily briefing tool.
4. Newsom hints at support for high-profile AI bill šļø
According to POLITICO, California Gov. Gavin Newsom signaled backing for AI safety efforts while name-checking leading academics, a remark widely read as a nod to state Sen. Scott Wienerās closely watched SB 53.
The bill would require some AI developers to publicly disclose safety and security practices, set up a mechanism to report major AI safety incidents to California, and protect whistleblowers, while laying groundwork for a state cloud cluster called CalCompute. After vetoing broader measures last year, Newsomās new tone suggests a friendlier path for transparency-focused regulation that the industry and safety advocates are lobbying hard on.
5. Perplexity opens up its massive Search API to try and challenge Google š„
Perplexity AI launched a real-time Search API that gives developers direct access to its web index spanning hundreds of billions of pages, positioning the startup as a fresh challenger to Googleās grip on search infrastructure, according to Perplexity.
The API touts sub-document precision, hybrid retrieval and structured, citation-rich outputs tailored for AI apps, which could help teams build smarter products without leaning on increasingly restricted Google or Bing pipelines. Developers can now tap Perplexityās real-time, citation-rich web index through a new Search API, offering a viable alternative to Google and Bing for building AI-powered products.
6. Gemini 2.5 Flash gets sharper formatting, reasoning, and image skills šø
Google rolled out upgrades to Gemini 2.5 Flash in the app that tighten response organization with headers, lists, and tables, boost step-by-step reasoning for homework help, and enhance image understanding for tasks like turning photographed notes into flashcards.
The update is live via the model picker, while new preview builds for developers in AI Studio and Vertex AI claim better tool use, faster outputs, and a 2x improvement in token efficiency for lower cost.
š¦¾How You Can Leverage:
You know those AI hackathons you hear about in San Francisco? š¤
Guess whoās cleaning house?
Paige Bailey told us the secret: non-coders are winning.
(Yeah, non-coders are routinely winning AI hackathons that have pretty much always requiredā¦. Coding.)
Paige is the AI Developer Relations Lead for Google DeepMind and she kinda blew our mind on todayās Everyday AI show.
Yeah, she showed us a trick or three about Googleās viral āNano Bananaā AI image tool.
But, she also demoed live how to build full stack apps from scratch in 2 minutes, how to use AI in Google Sheets to save hours and simplified it all even for non-technical people.
Sheeeeesh!
So, letās dive in to what you need to know.
1 ā Google Sheets AI Function Eliminates Research Teams š„
Your team manually researches addresses and supplementary data for business spreadsheets.
So 2023.
Paige wowed by pulling complete address information for 20 English soccer teams using a single function in Google Sheets. She typed equals AI, described what she wanted in plain English, and Gemini populated the entire column instantly with accurate location data.
This eliminates countless hours of internet research.
The breakthrough?
Fuzzy data cleanup that previously required Python programming knowledge. Misspelled city names get corrected automatically. Inconsistent capitalization gets standardized. Data preprocessing headaches that used to require technical teams now happen with natural language commands.
Paige emphasized this solves data problems that didn't even have exact functions before. Smart executives are using this for automated sentiment analysis on customer feedback while competitors manually categorize responses.
Try This
Think about any spreadsheet where you're missing information that requires manual research. Customer addresses, company details, contact information.
Open Google Sheets and type equals AI in an empty column. Like this: =AI() function
Describe exactly what missing information you need, like return the full corporate address for this company name, then reference the cell containing your data.
Watch Gemini in Sheets automatically populate your entire dataset. Guess you can go browse AI slop now with all that time you saved?
2 ā Natural Language Creates Deployable Apps With Nano Banana ā”
Non-technical people can't build functional web applications with cloud infrastructure.
Withā¦.. simple language. All in Googleās AI Studio with the āBuildā feature.
Paige built a webcam-enabled D&D character generator using only conversational descriptions in AI Studio's Build feature.
She described the app in natural language, and the system generated webcam integration, image processing with Nano Banana, character stats, professional UI design, and Google Cloud deployment.
And then it justā¦. Worked.
WTFmindblownemoji.
Weekend hackathons prove this democratization is real.
Try This
Think about a workflow problem your team encounters daily that could be solved with a simple application.
Access AI Studio and click the Build feature. Describe your business problem using normal conversation, like create an app that analyzes customer support tickets and ranks them by urgency for faster response times.
Watch the real-time progress as the system builds your complete application. When finished, deploy automatically with a shareable URL that your team can use immediately.
Test this thinking pattern: instead of asking how to code a solution, ask how to describe the problem clearly enough for AI to solve it.
3 ā Technical Barriers Are Disappearing for Business Teams š
Your company still separates technical and non-technical roles when building solutions?
How wildly pre-Gen AI of you. lolz.
Theyāre salespeople who spent ten thousand hours with customers, product managers with hobby projects, and business development people who deeply understand problems but never learned software engineering.
The competitive advantage shifted from technical implementation to problem articulation.
In other words, if you can accurate communicate it, Googleās Gemini can very likely build the first iteration.
Smart teams use Nano Banana for professional headshots because traditional photography isn't practical. They colorize historic photos instantly. They convert satellite imagery into game-ready assets that look professionally designed.
Five billion images created in under 30 days proves business teams discovered practical applications beyond creative experimentation.
Your marketing department could generate branded assets and professional imagery while competitors wait weeks for contractor deliverables.
Try This
Think about problems your team outsources because they seem too technical to solve internally. Or creative projects that could scale your biz that seemed too expensive or out of reach.
Then, identify one workflow that requires external tech or creative vendors that your team doesn't have budget for. Visual content creation, data analysis, simple application developmentā¦ you know.
Then, test whether you can build a first version using natural language descriptions in Google's AI tools instead of hiring specialists or waiting for development resources.
Shift your team's thinking from we need technical people to solve this toward we need to describe this problem clearly enough for AI to solve it.
Reply