Here’s a new episode of The Reddy Rundown, crafted so you don’t have to frantically follow everything in the AI news space wondering what you are missing as a an exec in 2025 trying to keep up

I am Shawn Reddy. CEO and Founder at AI Advantage Agency. I think like a systems first marketer who cares about creative that actually moves revenue. This week the signal is simple. New models change what we can prototype in minutes. Open weights change your cost profile and control surface. Agent upgrades change who on your team can ship.

Google’s Genie 3

What it is A world model that turns a single prompt into an interactive environment you can play with in real time. Think 24 fps scenes with memory and physics that react to you.

Why marketers should care

• Interactive demos without a game studio. Build a product walk through that users can explore instead of watch

• Rapid concept testing. Put two creative directions into play and watch where users go and where they stall

• Synthetic user sessions for training and QA. Feed your agents traces from controlled worlds before you throw them at live traffic

My take This is the bridge between static ads and experiential ads that actually teach. If you sell anything complex, plan on pilots where a prospect can try the idea not just see it.

OpenAI’s gpt oss 120B and 20B

What it is Open weight reasoning models under Apache 2.0 that run locally or in your cloud and compete with the lighter closed models.

Why marketers should care

• Lower unit costs at scale with real control over data and latency

• Private reasoning for campaigns that use customer data or partner SKUs

• Cleaner vendor risk. You can pin a model version and keep shipping while the rest of the world waits for an API change

My take If you run volume outreach, support, or personalization, this lets you move critical paths off rented brains. Expect your finance team to like the margin story.

NotebookLM Video Overview

What it is Drop documents in and auto build narrated videos with slides pulled from the source.

Why marketers should care

• Turn playbooks and case studies into quick buyer education assets

• Create role based versions for sales, success, and partners without a video team

• Speed up content atomization. One whitepaper becomes a LinkedIn carousel, a training clip, and a landing page explainer before lunch

My take Education beats interruption. This is an instant way to raise comprehension across your funnel without adding headcount.

Anthropic Claude Opus 4.1

What it is An incremental but real bump on coding, research, and multi step reasoning with better performance on messy real projects.

Why marketers should care

• Refactors your internal tools and scrapers faster

• Fewer silent errors in long running agent flows

• Better synthesis across decks, docs, and data when you need an answer not a paragraph

My take I care less about the benchmark digits and more about reliability. This update reduces the number of times I need a human to clean up after an agent.

Sōkosumi agent marketplace

What it is A pay per task marketplace for multi model agents with enterprise controls.

Why marketers should care

• Spin up temporary agents for research, creative variations, and data cleanup without adding subscriptions

• MCP ready makes orchestration sane inside your existing stack

• Good way to trial new workflows before you commit build time

My take Use this like a staffing bench for machines. Great for overflow and experiments.

What we are changing in our stack this week

• Experiential ad pilots with Genie style worlds for complex offers and onboarding

• Shift certain enrichment and summarization to gpt oss for cost control and privacy

• Adopt Video Overview as the default step in sales enablement and product marketing

• Move long running QA and code refactors to Opus 4.1 where stability matters

• Use Sōkosumi for bursty research and competitive scrapes instead of hiring contractors

Tools and how to use them

• Google Genie 3 Use case Build a playable product demo or onboarding mini world for education and pre qualification

• OpenAI gpt oss 120B Use case Run on your infrastructure for large scale outreach drafting, support macros, and rules based personalization with lower cost

• OpenAI gpt oss 20B Use case Laptop friendly model for field teams and analysts to run local research and quick data hygiene

• NotebookLM Video Overview Use case Turn docs into short explainers for sales and training then clip for social and email nurtures

• Anthropic Claude Opus 4.1 Use case Agent of record for code refactors, data audits, and deep research where accuracy matters

• Sōkosumi Use case Spin up pay per task agents for competitive research, creative variation sets, and spreadsheet cleanup

• Lovart Use case Co create brand safe visuals and maintain look and feel across campaigns with memory and style control

The bottom line

I am a systems first marketer turned CEO. These releases are not just cool. They change the math on time to first test, on cost per assisted click, and on how reliably an agent can run without a human babysitter. If you run growth, pick two pilots from the list above and put them into your plan this month. Then measure lift on assisted clicks, time to first meeting, and support resolution speed. If the numbers move, scale.

Looking for a community of like-minded individuals who are interested in AI and Entrepreneurship? Join our free community here to get started: The AI Advantage Community Thank you for reading! -Shawn