Google I/O

Summary

Tech Brew Ride Home – "Google I/O" — May 20, 2026

Overview

This episode dives deep into Google’s 2026 I/O keynote and the sweeping AI advances just announced. Host Brian McCullough and his co-host break down the headline releases—Gemini 3.5 Flash, Gemini Spark (the always-on AI agent), Gemini Omni (Google’s new multimodal “can do anything” model), a dramatic revamp of Google Search leveraging these models, updates to the agentic coding platform Anti Gravity, and creative tools like Project Genie. If you want a fast but thorough tour of what Google just unleashed—and what it means for the next AI chapter—this summary will get you there.

---

Key Discussion Points and Insights

1. Gemini 3.5 Flash: Google’s New Powerhouse Agentic Model

[02:10 – 05:25]

• What’s New:
  Google launched Gemini 3.5 Flash, described as their most powerful model yet for agentic and long-horizon tasks—AI that can autonomously manage multi-step workflows.
• Performance:
  • Flash is “four times faster than competing frontier models and that an optimized version pushes that figure to 12 times faster with equivalent quality.” (03:12–03:20)
  • Designed for speed in “agentic” environments with many agents working simultaneously.
  • Co-developed with “Anti Gravity,” Google’s agentic development platform and IDE.
• How It Works:
  • Runs autonomously for hours, only pausing when human judgment is required:

    “The model is built for independ[ence], but it is also built to know when to stop and ask.” (03:54–03:56, Brian McCullough)

  • Not a monolithic supermodel:

    “Google is not building a single all powerful model, it is building a hierarchy of specialists...” (04:11, Brian McCullough)

• Real Applications:
  • Banks and fintechs automating weeks-long workflows.
  • Data scientists surfacing insights without manual work.
  • Flash as the core AI in Gemini app/search; also powers Gemini Spark.
• Strategic Shift:
  Google repositioning from chatbot to agent builder:

  “The moment its AI stopped being a chatbot and started being a builder.” (02:43, Brian McCullough)

---

2. Gemini Spark: The 24/7 Personal AI Agent

[05:35 – 07:22]

• Headline:

  “Google has announced a 24.7 personal AI agent called Gemini Spark at this year's IO Developer Conference.” (05:47, Tech Brew Co-host quoting Engadget)

• Integration:
  • Deeply connected to Gmail, Docs, Slides, and now third-party apps like Canva, OpenTable, and Instacart.
  • Performs multi-step, recurring, and cross-app workflows for users.
• Capabilities:
  • Summarize long email threads, spot hidden fees, auto-generate and send reports, or compile to-do lists.
• Control & Safety:
  • Spark is opt-in and users choose which apps it can access.
  • Before any “high stakes” action (like sending emails or spending money), Spark will ask for permission.
• Rollout:
  • Rolling out first to testers, then AI Ultra Beta users in the US. Desktop integration coming this summer.
  • Future features will allow Spark to send texts/emails and operate your browser.
• Vision:

  “Transforms Gemini from a standard AI assistant to an active partner that actually performs tasks for you.” (05:54, Brian McCullough quoting Engadget)

---

3. Gemini Omni: Multimodal Model for Video and More

[07:31 – 11:41]

• Announcement:
  • Gemini Omni is a natively multimodal model:

    “Can create anything from any input, starting with video generation.” (07:31, Brian McCullough)

  • Accepts text, images, audio, video as inputs, produces outputs across the same.
  • Builds on Google’s past with Nano Banana, but now all-in-one instead of modular pipelines.
• Technical Leap:

  “A unified model can reason across modalities in the same forward pass, which generally translates into more coherent edits, fewer pipeline artifacts, and a far cleaner API surface for developers.” (08:05–08:15, Tech Brew Co-host)

• Interaction:
  • Conversational video editing—each instruction refines the output, state persists, enabling complex iterative edits.
  • Improved physics: “gravity, kinetic energy, fluid dynamics … separates looks like AI video from looks like real footage.” (09:17, Tech Brew Co-host)
• Enterprise Applications:
  • Uses: rapid generation of localized ads, explainers, onboarding, visual product demos, dynamic support videos, simulations, and more—not just creative arts.
• Rollout & Availability:
  • Omni Flash available in Gemini app (paid AI tiers), API to Vertex AI coming soon.
• Content Safety:
  • All Omni videos are digitally watermarked (SynthID), expanded C2PA credentials, launch of content detection API for businesses.
  • Major partners: Shutterstock, Avid, major newswires onboard.
• Market Positioning:
  • Competes with OpenAI’s earlier GPT-4 Omni, but with emphasis on video and enterprise-readiness.

---

4. Google Search Overhaul: AI at the Core

[13:21 – 15:36]

• Major Update:
  • First redesign of the Google search box since 2001.
  • Now supports longer queries, uploads of images/video, and persistent chat-style follow-ups.

  “For the first time since 2001, the box is getting bigger and more interactive … people can ask follow up questions with a chatbot on Google's main search page.” (13:36–13:51, Tech Brew Co-host)

• AI-Driven Features:
  • Powered by Gemini 3.5 Flash, transforming search into an agentic, conversation-like experience.
  • Digital assistants (agents) can automate ongoing searches, e.g. notify you of new apartment listings.
  • AI Overviews (summaries instead of website lists) and an “AI Mode” search tab are being combined into one seamless interface.
• Impact on Business & Content:
  • Noted shift in how Google monetizes:

    “Google is reducing everyone to raw data providers.” (14:50, Brian McCullough attributing Richard Kramer of Arete Research)

  • Queries on complex topics now return interactive graphics/simulations powered by Gemini code, not just links.

---

5. Google AntiGravity 2.0: Agentic Coding Gets a Boost

[15:46 – 17:24]

• Update:
  • New desktop app, CLI tool, SDK for building custom agent workflows.
  • Users can orchestrate multiple coding agents and automate/schedule workflows.

  “With the new desktop app, users can orchestrate multiple agents and execute tasks simultaneously.” (16:08, Brian McCullough)

• Integration:
  • Works with Google AI Studio, Android, and Firebase.
  • Voice command support and a new terminal CLI for developers.
  • Gemini CLI users must migrate to new tool.
  • SDK and pre-built templates aid enterprise integration.
• Underlying Engine:
  • All driven by Gemini 3.5 Flash and Anti Gravity’s architecture.
• Vision:
  • Makes agentic software development mainstream—“agent is the name of the game right now.” (05:35, Brian McCullough)

---

6. Project Genie: AI-Powered 3D World Generation with Street View

[17:33 – 19:08]

• Expansion:
  • Project Genie rolls out to all Google AI Ultra users, now with Street View photo integration.
  • Users can base AI-generated environments on real-world locations from Google Maps (starting with the US).
• How It Works:
  • Generates 60-second 3D “snow globe” scenes in 720p, using written prompts and customized avatars.
  • Camera can explore space; not a full-fledged game maker.

  “Project Genie is not a game generator … Project GENIE is capable of generating a limited 3D environment with a controllable camera attached to an avatar.” (18:40–19:08, Brian McCullough)

• Example:
  • Demo: Golden Gate Bridge transformed into an underwater scuba scene.

---

Notable Quotes & Memorable Moments

• Turning Point for AI:

  “The moment its AI stopped being a chatbot and started being a builder.” (02:43, Brian McCullough)

• Speed Emphasis:

  “Flash outperforms our latest Frontier Model 3.1 Pro on nearly all benchmarks, including coding, agentic tasks and multimodal reasoning.” (03:02, Brian McCullough quoting Corai Kavuciculo, DeepMind)

• Agent Model Philosophy:

  “Google is not building a single all powerful model, it is building a hierarchy of specialists ...” (04:11, Brian McCullough)

• The New Search:

  “For the first time since 2001, the box is getting bigger and more interactive so that people can ask even longer questions and upload photographs and videos into queries.” (13:36, Tech Brew Co-host)

• Open Web's Future:

  “The open web is on its way out, Mr. Kramer said, referring to the way Internet traffic now often begins and ends with a visit to Google.” (14:41–14:44, Brian McCullough attributing Richard Kramer)

---

Timestamps for Key Segments

• Gemini 3.5 Flash & Agentic Models: 02:10–05:25
• Gemini Spark (24/7 Personal Agent): 05:35–07:22
• Gemini Omni (Multimodal Model): 07:31–11:41
• Google Search Overhaul: 13:21–15:36
• Anti Gravity 2.0 (Agentic Coding): 15:46–17:24
• Project Genie & Street View Worlds: 17:33–19:08

---

Summary

Google’s 2026 I/O was a landmark for pushing agentic, always-on AI to the heart of consumer and enterprise tech. Gemini 3.5 Flash brings agent speed and autonomy; Spark acts as an AI task partner across apps and third-parties; Gemini Omni sets a new multimedia standard; and Search, the core of Google’s business, is now an interactive AI experience, not a list of links. Coding agents via Anti Gravity enter the mainstream, and creative possibilities expand with tools like Project Genie. Underlying it all: the shift from AI as a button-pushing assistant to AI as an initiator, orchestrator, and builder across the digital world.

---

(End of summary. All ads, non-content sections, and unrelated conversation have been omitted for clarity and flow.)

Transcript

[Windows 11 Student Offer Announcer]

Study and play come together on a Windows 11 PC and for a limited time, college students get the best of both worlds. Get the unreal college deal everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 Premium and a year of Xbox Game Pass ultimate with a custom color Xbox wireless controller. Learn [e-mail] studentoffer while supplies last ends June 30th terms at aka mscollegepc.

[Brian McCullough]

Welcome to the Tech Brew Ride home for Wednesday, May 20, 2026. I'm Brian McCullough. Today, Google I O was Gemini 3.5

[Tech Brew Co-host]

Flash, its fastest agentic model yet, Gemini Spark as a 24. 7 personal agent.

[Brian McCullough]

Another new model, Gemini Omni for video generation, an AI overhaul to the search box and Street View is coming to Project genie.

[Tech Brew Co-host]

Here's what you missed today in the world of tech.

[Brian McCullough]

Today's episode is brought to you by Doppl.

[Tech Brew Co-host]

Social engineering attacks don't even bother to knock. They slip right into your inbox, your

[Brian McCullough]

phone, or on websites, instead pretending to be a harmless internal email or a normal text message until it's too late. Doppel sees right through these disguises. Their AI native platform trains your team

[Tech Brew Co-host]

to recognize threats like deepfakes, bad links, and impersonation attempts before they can actually cause any damage.

[Brian McCullough]

Doppel strengthens team resilience by giving employees the tools and defenses they need to

[Tech Brew Co-host]

protect themselves from increasingly sophisticated social engineering threats.

[Brian McCullough]

It's kind of like having a security

[Tech Brew Co-host]

team that has eyes everywhere, and their digital risk protection takes it one step

[Brian McCullough]

further by keeping an eye on every

[Tech Brew Co-host]

channel to connect patterns and shut them down fast. Invest in social engineering defense today. Learn more at the-o P E L.com

[Brian McCullough]

that's-o P P E L.com well, as promised, Google announced a ton of stuff

[Tech Brew Co-host]

at I O yesterday, and there's basically no way we can talk about everything that was announced. So here is everything I think matters.

[Brian McCullough]

First up, Google rolled out Gemini 3.5

[Tech Brew Co-host]

flash, its quote strongest agentic encoding model yet for tackling long horizon agentic tasks

[Brian McCullough]

in the Gemini app and searches AI

[Tech Brew Co-host]

mode quoting the Next Web Somewhere inside

[Brian McCullough]

Google's campus, an AI agent recently built an operating system from scratch. Not a toy demo, not a sanitized benchmark exercise, but a full OS assembled by software agents spawning off to handle

[Tech Brew Co-host]

separate components before stitching everything together.

[Brian McCullough]

That, at least, is what Google showed

[Tech Brew Co-host]

the world at I O 2026 on

[Brian McCullough]

Monday, and it wants the demonstration to mark a turning point. The moment its AI stopped being a chatbot and started being a builder. The model behind the feed is Gemini 3.5 Flash, which Google is calling its strongest model yet for coding and autonomous AI agents.

[Tech Brew Co-host]

According to Corai Kavuciculo, DeepMind's chief technologist,

[Brian McCullough]

Flash outperforms our latest Frontier Model 3.1 Pro on nearly all benchmarks, including coding, agentic tasks and multimodal reasoning.

[Tech Brew Co-host]

The company says the model is four

[Brian McCullough]

times faster than competing frontier models and that an optimized version pushes that figure to 12 times faster with equivalent quality.

[Tech Brew Co-host]

Speed is the point when agentic AI systems run in production, they do not take turns. Multiple agents operate simultaneously on long running

[Brian McCullough]

tasks, and latency becomes a bottleneck that

[Tech Brew Co-host]

compounds with every parallel process. Google has built Flash specifically for that

[Brian McCullough]

environment, co developing it alongside Anti Gravity,

[Tech Brew Co-host]

the company's agentic development platform and ide. Flash can run autonomously for multiple hours,

[Brian McCullough]

according to Google, pausing to request user input only when it reaches a decision

[Tech Brew Co-host]

point or permission boundary that requires human judgment. That design sits in deliberate tension. The model is built for independ, but

[Brian McCullough]

it is also built to know when

[Tech Brew Co-host]

to stop and ask.

[Brian McCullough]

Tulsi Doshi, Google's senior director and head

[Tech Brew Co-host]

of product for the Gemini model, outlined the broader architecture.

[Brian McCullough]

A forthcoming 3.5 Pro model will serve as the orchestrator and planner while Flash handles the sub agent work.

[Tech Brew Co-host]

The division is instructive.

[Brian McCullough]

Google is not building a single all powerful model, it is building a hierarchy of specialists with pro thinking and Flash

[Tech Brew Co-host]

doing partners are already testing the approach. Banks and fintechs are reportedly automating multi

[Brian McCullough]

week workflows and data science teams are

[Tech Brew Co-host]

using Flash agents to surface insights that previously required manual investigation.

[Brian McCullough]

Google has not disclosed specific client names

[Tech Brew Co-host]

or quantified results, so those claims remain difficult to independently verify the launch signals.

[Brian McCullough]

A broader strategic pivot Google is no longer positioning its AI primarily as a conversational interface. Flash is the default model in the

[Tech Brew Co-host]

Gemini app and powers AI mode in search globally.

[Brian McCullough]

It also underpins Gemini Spark, the new

[Tech Brew Co-host]

personal AI agent that runs around the clock on Google cloud infrastructure, handling tasks across Gmail, docs and more than 30 third party integrations.

[Brian McCullough]

The timing matters. Vibe coding and AI assisted development have moved from novelty to production tooling in the span of a year, and every major lab is racing to claim the agentic layer.

[Tech Brew Co-host]

OpenAI, anthropic and a growing roster of

[Brian McCullough]

startups are building competing agent frameworks. Google's bet is that raw speed, tight

[Tech Brew Co-host]

integration with its own cloud infrastructure and

[Brian McCullough]

purpose built IDE will give Flash an

[Tech Brew Co-host]

edge that benchmarks alone cannot capture End quote.

[Brian McCullough]

Agent is the name of the game

[Tech Brew Co-host]

right now, of course, and Google wants Gemini Spark to be your personal agent. Quoting Engadget Google has announced a 24.7personal

[Brian McCullough]

AI agent called Gemini Spark at this year's IO Developer Conference.

[Tech Brew Co-host]

The company says Spark transforms Gemini from

[Brian McCullough]

a standard AI assistant to an active partner that actually performs tasks for you.

[Tech Brew Co-host]

Spark is powered by Gemini 3.5 and

[Brian McCullough]

is deeply integrated with Google Workspace apps including Gmail, Docs and Slides.

[Tech Brew Co-host]

You can teach it to perform various tasks, such as creating a list of critical deadlines in your Gmail and sending

[Brian McCullough]

it to you, or writing up a

[Tech Brew Co-host]

summary of ongoing updates in lengthy email threads.

[Brian McCullough]

It's also possible to set it to do recurring tasks like making it spot hidden fees and credit card bills every month.

[Tech Brew Co-host]

You can program it to to complete several interconnected tasks for complete workflows as well.

[Brian McCullough]

For instance, you can ask Spark to look at meeting notes in your chats

[Tech Brew Co-host]

and emails and create polished reports in Google Docs, as well as draft an email that you can send along with that report. Gemini Spark is rolling out to testers

[Brian McCullough]

shortly before making its way to Google AI Ultra Beta users in the US next week.

[Tech Brew Co-host]

Spark is completely opt in.

[Brian McCullough]

It's up to you whether to turn it on and use it, and you

[Tech Brew Co-host]

can also choose which apps it can connect to.

[Brian McCullough]

In addition to workspace apps, it will be able to link to Canva, OpenTable and Instacart right now, with more partner

[Tech Brew Co-host]

apps coming in the following weeks.

[Brian McCullough]

Google will also release more features for the AI agent in the coming weeks,

[Tech Brew Co-host]

including being able to give it the

[Brian McCullough]

ability to send texts and emails, and the ability to operate your browser.

[Tech Brew Co-host]

The company assures that Spark will ask you first before performing high stakes actions like spending money or sending emails.

[Brian McCullough]

Google also says that it's bringing Spark to the Gemini desktop app this summer so that it can access files and

[Tech Brew Co-host]

perform tasks on your computer.

[Brian McCullough]

Gemini Omni is a new multimodal model Google says can create anything from any input, starting with video generation.

[Tech Brew Co-host]

Quoting VentureBeat, Omni is the next chapter

[Brian McCullough]

of the work that produced Nano Banana,

[Tech Brew Co-host]

the image generation and editing model Google shipped roughly a year ago.

[Brian McCullough]

The first model in the family, Gemini Omni Flash accepts any combination of text, images, audio and video as images input and produces high quality output across the same modalities, all from a single model

[Tech Brew Co-host]

rather than a relay of specialized systems.

[Brian McCullough]

Google says the model is natively multimodal

[Tech Brew Co-host]

from the ground up, which matters less as marketing company than as an architectural claim. A unified model can reason across modalities

[Brian McCullough]

in the same forward pass, which generally

[Tech Brew Co-host]

translates into more coherent edits, fewer pipeline artifacts, and a far cleaner API surface for developers.

[Brian McCullough]

OpenAI started this trend back in May

[Tech Brew Co-host]

2024 with the release of GPT4, its first natively Omni model also trained from the ground up to analyze and generate multiple different types of content, from text

[Brian McCullough]

to code, imagery and audio. However, it did not support video generation and the model was eventually deprecated following reports of sycophancy and Even users demanding

[Tech Brew Co-host]

OpenAI retain it after developing parasocial relationships with it.

[Brian McCullough]

Is Gemini Omni at risk of sparking

[Tech Brew Co-host]

a similarly devoted following? It remains to be seen.

[Brian McCullough]

One big difference though, is that its headline interaction pattern is conversational video editing. Each instruction builds on the last and past directions persist across turns, so the video evolves coherently as the user iterates.

[Tech Brew Co-host]

Practical examples Google highlighted include changing the world inside a clip, reimagining an action

[Brian McCullough]

or camera angle, refining techniques over multiple

[Tech Brew Co-host]

turns, and generating explainer style content from short prompts. Google also emphasizes improved physics, gravity, kinetic energy, fluid dynamics, which is the kind of detail that separates looks like AI video from looks like real footage. The first thing enterprise leaders should read

[Brian McCullough]

carefully is the rollout plan.

[Tech Brew Co-host]

Omniflash is going live today inside the gemini app for US subscribers across AI, AI Pro and AI Ultra tiers, including

[Brian McCullough]

the new $100 per month AI Ultra

[Tech Brew Co-host]

plan Google announced at the same event.

[Brian McCullough]

Google says it will roll out to developers via Vertex AI APIs in the coming weeks.

[Tech Brew Co-host]

That gap is significant until the Vertex API is generally available. Omni is effectively a consumer and prosumer tool. Enterprise pilots beyond individual seat based experimentation should wait for the API, both because that's where Google's enterprise SLAs and data

[Brian McCullough]

handling commitments live, and because production grade

[Tech Brew Co-host]

generative video without a programming interface is a non starter.

[Brian McCullough]

Its pricing through the API per million

[Tech Brew Co-host]

tokens presumably will also determine its viability as an enterprise product outside of film, tv, entertainment and the arts productions.

[Brian McCullough]

It's easy to default to marketing video as the use case, but but Omni's value proposition for enterprises is broader. If you think of it as a programmable video and media engine rather than

[Tech Brew Co-host]

a creative app, think of sales and marketing rapid generation of variant ads, localized creative and product demos with per asset

[Brian McCullough]

agency cycles, internal communications, learning and development would mean explainer videos on boarding modules

[Tech Brew Co-host]

and policy walkthroughs produced by non specialists. Think of customer support and documentation, dynamic

[Brian McCullough]

query conditioned visual explainers attached to help articles, product and engineering visualizations of simulations, UI walkthroughs and concept videos for spec reviews. Finally, think of field operations short situation specific instructional clips generated on demand for CIOs and CISOs. The most important section of Google's announcement is not the model card, it's the provenance and content safety work shipping alongside it. Every video generated by Omni carries Google's Synth ID digital watermark. Google is expanding C2PA content credentials across its generative tools and launching an AI content detection API on Agent platform that lets businesses identify AI generated content from

[Tech Brew Co-host]

both Google and other popular models. Partner integrations announced at the same event,

[Brian McCullough]

including Shutterstock, Avid in Pro Tools, and at least one major newswire, indicate where

[Tech Brew Co-host]

the standard is going.

[Brian McCullough]

Father's Day is right around the corner. Celebrate the dads in your life with Viori. Their Father's Day gift guide features the newest and best selling styles dad will want to wear on repeat. From the fair to the airport to the gym, Vuory's everyday staples are ready for wherever the day takes him. With Fitness, Go to's Performance Essentials and Travel Ready layers, you can gear him up for his next adventure in one place. Get 20% off your first order at viori.com techbrew that's vuori.com Techbrew exclusions apply. Visit the website for full terms and conditions. If your day is filled with discussion, ideas and constant back and forth, you don't always have time to take notes. That's where things get lost, and that's where Plaud Note PIN S comes in. It's a small, wearable AI device that's

[Tech Brew Co-host]

always ready so you can capture conversations

[Brian McCullough]

and ideas completely hands free. No need to unlock your phone or break your flow. Just tap to record and keep going. Plaud turns everything into clean notes, summaries and key takeaways automatically. It's also built with privacy in mind with enterprise grade compliance like SoC2, HIPAA and GDPR. Stay present in the moment without losing what matters. Check it out at plaud AI tbrh that's plaud p l a u d AI tbrh use code tbrh for 10%

[Tech Brew Co-host]

off all plod devices

[Brian McCullough]

all of this AI stuff is continuing to trickle down, cascade down to the main thing at Google. Google's Search BO on the Web Quoting the Times on Tuesday, Google said the AI shift had inspired it to overhaul the dimensions of its search box.

[Tech Brew Co-host]

For the first time since 2001, the

[Brian McCullough]

box is getting bigger and more interactive so that people can ask even longer questions and upload photographs and videos into queries. In addition, people can ask follow up questions with a chatbot on Google's main search page.

[Tech Brew Co-host]

The company will also offer digital assistants, known as agents, to automate searches so

[Brian McCullough]

that someone who might be apartment hunting can be notified of a new listing

[Tech Brew Co-host]

without opening a real estate site like Zillow.

[Brian McCullough]

The search features will be powered by a new artificial intelligence model, Gemini 3.5 Flash.

[Tech Brew Co-host]

Google said the model had improved on creating software code and performing autonomous tasks, worked faster and was less expensive to

[Brian McCullough]

run than comparable models. Sundar Pichai, Google's chief executive, said Gemini's speed and affordability made it possible to deliver it broadly, which will ultimately benefit Google. Richard Kramer, a financial analyst with Arete Research, said the changes were helping Google make more money from advertising. Last year, Google's ad clicks rose 6% and it charged 7% more for each click. The company's annual profit has more than

[Tech Brew Co-host]

doubled since 2022, to $132 billion.

[Brian McCullough]

The open web is on its way out, Mr. Kramer said, referring to the

[Tech Brew Co-host]

way Internet traffic now often begins and ends with a visit to Google. Rather than visiting other sites with AI,

[Brian McCullough]

Google is reducing everyone to raw data providers. Google's AI transformation predicting particularly stands out with search. In 2024, the company stopped fulfilling some queries with a list of websites and

[Tech Brew Co-host]

instead provided automatically generated responses called AI Overviews.

[Brian McCullough]

Last year, it added a search tab

[Tech Brew Co-host]

called AI Mode, where people can ask multiple questions on the same subject as they would with a chatbot.

[Brian McCullough]

Google said those features were being combined. Now, on searches that deliver AI overviews,

[Tech Brew Co-host]

people can ask follow up questions in AI mode, which Mr. Pichai called a revelation.

[Brian McCullough]

Google is also bringing one of AI's biggest breakthroughs, software COD, to search. When people research complex topics like astrophysics, Gemini can build interactive graphics and simulations behind the scenes to provide a deeper

[Tech Brew Co-host]

answer than its previous listing of websites. End quote.

[Brian McCullough]

And don't forget that agentic Coding itself,

[Tech Brew Co-host]

quoting TechCrunch at Google I O 2026,

[Brian McCullough]

the company unveiled a new version of its agentic coding app app Google AntiGravity 2.0 with an updated desktop app, a CLI tool and an SDK for custom workflows.

[Tech Brew Co-host]

The company launched its Anti Gravity tool last year as a response to agentic coding software such as Cursor.

[Brian McCullough]

The company said that with the new desktop app, users can orchestrate multiple agents and execute tasks simultaneously.

[Tech Brew Co-host]

Plus you can design custom sub agent

[Brian McCullough]

workflows and schedule tasks that can automatically

[Tech Brew Co-host]

run in the background.

[Brian McCullough]

The new app also easily lets you integrate projects with Google, AI Studio, Android and Firebase.

[Tech Brew Co-host]

A lot of this is powered by by the company's new Gemini 3.5 flash model, which was co developed using Anti Gravity, Google said.

[Brian McCullough]

Google is also adding native voice command support to Anti Gravity just like it

[Tech Brew Co-host]

has added to multiple consumer products including Gmail and Docs. And it's launching a new Anti Gravity

[Brian McCullough]

CLI command line interface tool for programmers who want to use a terminal for creating agents. Google is asking users who use the

[Tech Brew Co-host]

Gemini CLI tool to migrate to the new tool. What's more, Google is launching an Anti

[Brian McCullough]

Gravity SDK for developers to build custom

[Tech Brew Co-host]

agents based on Google's coding tools. The company will allow Google Cloud customers to connect to Anti Gravity to build projects. It noted that it will release custom

[Brian McCullough]

agent templates in AI Studio for enterprise users to get started. What's more, Google is launching an Anti Gravity SDK for developers to build custom agents based on Google's coding tool.

[Tech Brew Co-host]

The company will allow Google Cloud customers

[Brian McCullough]

to connect to Anti Gravity to build projects.

[Tech Brew Co-host]

It noted that it will release custom

[Brian McCullough]

agent templates in AI Studio for enterprise

[Tech Brew Co-host]

users to get started.

[Brian McCullough]

One last one Google has also added Street View integration to Project Genie.

[Tech Brew Co-host]

Quoting Engadget Project Genie is rolling out today for all adult Google AI Ultra

[Brian McCullough]

subscribers across the globe.

[Tech Brew Co-host]

Following its debut in the United States

[Brian McCullough]

this January, the service is also getting a new Street View capability that can generate interactive landscapes based on real world locations found on Google Maps, starting with

[Tech Brew Co-host]

places in the U.S. project Genie is

[Brian McCullough]

Google's AI powered system for creating explorable snow globe environments from written prompts with

[Tech Brew Co-host]

creations lasting 60 seconds at 720p and 24 frames per second. Users are able to create contained worlds

[Brian McCullough]

in whatever style they'd like, complete with

[Tech Brew Co-host]

a character of their own description, and

[Brian McCullough]

then move a camera around that space. The fresh Street View functionality allows users

[Tech Brew Co-host]

to base their AI worlds on location photos pulled from Google Maps, grounding their creations in a snapshot of reality.

[Brian McCullough]

In a demo video shown at I O, Google turned the Golden Gate Bridge into an underwater scuba scene, for instance.

[Tech Brew Co-host]

Google plans to expand the Street View interaction to more real world locations over time.

[Brian McCullough]

On the surface and in Google's ads, Project Genie looks like a cocoon world generator, producing perfect little video game bubbles

[Tech Brew Co-host]

out of vague written prompts.

[Brian McCullough]

But that's not really what's happening here. Project Genie is not a game generator by any means, and it's silly to

[Tech Brew Co-host]

suggest so among a suite of technical differences.

[Brian McCullough]

It ignores the expertise and precision required

[Tech Brew Co-host]

to create consistent, responsive and bespoke mechanisms

[Brian McCullough]

in persistent and believable worlds, and belies the importance of narrative flow in interactive stor stories. Project GENIE is capable of generating a limited 3D environment with a controllable camera attached to an avatar.

[Tech Brew Co-host]

This is cool, but it is not game development.

[Brian McCullough]

A couple of people have asked me

[Tech Brew Co-host]

why I've not been talking about Arsenal

[Brian McCullough]

in the last few weeks. Two reasons.

[Tech Brew Co-host]

Number one, I didn't want to bore people by talking about something they probably

[Brian McCullough]

don't care about over and over again.

[Tech Brew Co-host]

And believe me, I've been thinking about it over and over again. But I spared you.

[Brian McCullough]

But also, I legit didn't want to jinx it.

[Tech Brew Co-host]

So yes, if you're not aware, Arsenal claimed their first league title in 22 years yesterday. It was a very, very good day between myself and my kids. I have not probably in my life scrolled social media so voraciously as I have over the last 24 hours. I literally we cannot get enough Arsenal winning the title Content Talk to you tomorrow.

[Carrington College Announcer]

Your next chapter in healthcare starts at Carrington College's School of Nursing in Portland. Join us for our open house on Tuesday, January 13th from 4 to 7pm you'll tour our campus, see live demos, meet instructors, and learn about our Associate Degree in Nursing program that prepares you to become a registered nurse. Take the first step toward your nursing career. Save your spot now at Carrington. Edu Events. For information on program outcomes, visit carrington. Edu Sci Fi.