Spencer Forman

Is the browser your agent now? What happens when your web browser becomes an assistant? In this Stewart Squared episode, we explore the rise of agentic browsers like Atlas, Comet, and Gemini, and how they promise to change how we use the internet. We move from the old search bar to a world where your browser thinks, acts, and builds alongside you. It’s not science fiction, it’s the next user interface revolution. Watch the whole episode on @Stewart Squared 

17

I’ve been working on ConsentWatch for a while now, and seeing it live feels real. We kept running into the same problem: banners that don’t actually respect consent, tools collecting data people never agreed to, and businesses stuck between legal risks and broken setups. ConsentWatch is our answer to that mess. It gives users a way to report shady practices and gives businesses a way to see what’s really happening in their stack. What we do on the business side is straightforward: - Consent and Tracking Audits to check banners, trackers, and CMPs against GDPR, CCPA, and more - AI Data Privacy Audits to see how models collect, process, and store personal data of your end users. Both audits give clear actions, not 50-page reports nobody reads. This one matters to me because privacy shouldn’t be smoke and mirrors. It should be clear, fair, and accountable. #privacy #gdpr #ccpa #consentmanagement #aiprivacy #trust

3

I think a lot of people are sleeping on the progress Google is making in AI. Some of my favorite tools right now are: - Google AI Studio for full app design and development - Stitch for UI design - Antigravity as one of the best AI software development tools out there. Gemini has become my go-to for deep research, and NotebookLM is incredible.

33

2 Comments

Breeders need a trustable home for their life’s work, and new ways to monetize IP. Gentics, brands, and artwork locked into verifiable on‑chain contracts that protect authorship and automate royalty payments every time their work moves. AI that helps breeders predict and design new lines, while the agency and platform match them with licensing opportunities worldwide. Fans, growers, and partners able to invest alongside their favorite Breeder through tokenized deals and prediction markets. Community rails that tie it all together so everyone can see where the genetics came from, who gets paid, and who can be trusted. A community platform that serves as the global marketing, distribution and customer feedback loop engine. secure IP, codify reputation, get paid automatically, and connect to the people and capital that let Breeders keep creating.

15

17 Comments

Before you build an AI agent, answer this: Which workflow is costing you the most time right now? Missed calls. Unanswered estimates. Manual reminders. That’s where most people get it wrong. They start with tools instead of the problem. Here’s the 4-step way to launch your first AI agent in a home service business, without overcomplicating it: • Choose the single workflow creating the most drag • Define what should be said and when a human steps in • Connect it to the tools you already use • Monitor results, refine, then duplicate Build one agent that works. Prove it. Then scale it. Follow Tersh Blissett for practical AI systems that actually hold up.

21

15 Comments

“Even if AI firms are reined in and are made to only use works with permission or compensation...there will be significant pressure on copyright holders to allow AI training.” - FRANK LANDYMORE https://lnkd.in/gpBCM48H

3

1 Comment

Gamma + n8n + GPT-5 is kind of wild... It basically turned into an AI presentation engine that pulled in $47K in deck services last month. 🤯 While agencies charge $3K–$5K per deck and take two weeks… this thing spits out investor-grade presentations in *11 minutes.* No designers. No PowerPoint prison. No revision hell. 🙏 Here’s the workflow in plain words: → Client fills a form with 5 things: topic, audience, goal, brand colors, deadline → n8n fires off a GPT-5 run that builds the entire narrative architecture → Gamma renders a polished, investor-ready deck instantly → System auto-exports to Drive, drafts the client email, logs everything in Sheets → Every output is tracked so delivery is literally instant And the whole stack runs 24/7 while you sleep. 😴⚙️ Why it feels unfair: → You generate 50 variations while competitors build 1 manually → Each deck costs pennies, you charge $500–$2K → No designers slowing anything down → Quality looks Fortune-500 level → Delivery happens automatically through pre-written emails One consultant used this to productize their entire firm. They jumped from 4 custom decks a month to 40+. Same workload… just scaled by automation. 📈 The real shift is psychological: You’re not selling “presentations.” You’re selling *presentation infrastructure* that compounds.

1

I analyzed 8 million crime records and found a huge problem. We're building a model pipeline on the data from Chicago. That's because Chicago is super rich, much like other cities, New York, London, San Francisco, there is a set of publicly available crime data. The starting point is to break down the city into grids and then further that data breakdown into grid days. But there's a problem. 75% of the grid days have no crime. That makes it increasingly hard to train a model to predict a distribution of crime when it doesn't have any data to base those predictions on. But not to worry, we have a fix. Let's give the model the ability to predict nothing. That is no crime on that grid day. From a usefulness perspective, that is as useful to an end user as predicting crime. This gives us the ability to route traffic, to route delivery drivers, change the way you walk to work, etc. These updates are a couple days behind, so I'll give you a hint: this change made a huge difference. See you in the next one. 

11

2 Comments