💡 The $1.5B Bet That Says Agents Aren’t a Demo

In partnership with

☕ Morning! 💡 Your Weekly 5-Minutes of Caffeine and Tech Clarity

Quick Hits 🎯

• 🤖 Google turned Gemini into a coworker, not a chatbot. It can now plan and run multi-step tasks across your business apps

• 🧠 Nous Research hit a $1.5B valuation and shipped agents built for business users. The money is betting on agents, not demos

• 🎧 Spotify quietly opened an enterprise software arm. The consumer giant wants to sell you the tech stack it built to run itself

• 🧪 HackerRank built an AI that runs the technical interview. A preview of how hiring pipelines get standardized at scale

• 🤖 MIT researchers say humanoid robots are still a slow grind. Reliability, not imagination, is the real bottleneck

• 🧬 The scientist behind AI-designed viruses sat down to talk guardrails. Provenance and containment are becoming the actual product

• 🧩 Apple, LG, and Amazon are circling the same question: who sets the standard when every device ships its own agent?

🎁 + 6 other stories you might find useful (collaborator: attach links)

Our Partner 🎉 

Webinar Series: Running Agents in Customer Work

Every AI vendor reports a resolution rate. No two calculate it the same way, and none of them tell you whether the replies were actually good.

Running Agents in Customer Work is four 30-minute sessions for the leaders who have to make agentic AI work in support, success, and ops. Session 4 gets into how to score quality at volume without reading every transcript, and the questions to ask a vendor about what their resolution math leaves out.

Four Tuesdays, Oct. 13 through Nov. 3, 10 a.m. PT. Register today, one sign-up covers all four.

Register Now

The Big Picture 🖼️

💡 Google Just Turned Its Chatbot Into a Colleague.

Gemini is becoming an enterprise agent that can plan work, act across your apps, and spin up subagents with its own identity inside your workplace systems.

That’s a different thing from autocomplete. Once an AI gets a login and permission to act, the hard question stops being “is it smart?” and becomes “what can it touch, and who’s accountable when it moves?”

For everyone from intern to CTO, the skill that matters now is knowing where to point an agent and where to fence it off. Here’s what changes the day your tools start acting on their own.

💡 The Money Is Betting on Agents, Not Demos.

Nous Research confirmed a $1.5B valuation and launched agents aimed squarely at business users.

A number that size isn’t paying for a slick demo, it’s paying for the bet that companies will hand real workflows to agents and keep them running in production. The differentiator is already shifting from “can it do the task” to “can we govern it at scale.” That’s where the smart money thinks the moat actually is.

💡 Spotify Wants to Sell You Its Plumbing.

Spotify opened up an enterprise software arm, packaging the internal tech it built to run itself.

When a consumer company decides its backend is worth selling, it’s a signal worth reading: the real margins in software keep moving toward platforms and developer ecosystems, not features.

The interesting move is guessing which “consumer” brand quietly becomes an infrastructure vendor next. Here’s why a music company is suddenly in your procurement pipeline.

💡 The Job Interview Is Getting Automated From the Inside.

HackerRank shipped an AI that conducts the technical interview itself.

Consistency and scale are the pitch. Same questions, same scoring, fewer human biases in theory. The catch: whoever designs the rubric now shapes who gets hired across thousands of companies.

Whether you run interviews or sit through them, the game is changing underneath you. This is what hiring looks like when the interviewer never gets tired.

💡 The Robots Aren’t Coming as Fast as the Headlines Say.

One of the field’s founding figures. The roboticist behind the Roomba - just called the race to build human-shaped robots “pure fantasy thinking,” arguing today’s machines can’t learn real-world dexterity the way the hype assumes.

It’s the counterweight to agent mania. The near-term win isn’t a robot that does everything - it’s a narrow, dependable system you can actually explain and audit.

When a pioneer says the money is chasing the wrong shape, it’s worth asking which of your own “AI will handle it” bets are demos dressed up as products. Here’s why a Roomba pioneer says the humanoid bubble is set to burst.

💡 Governance Is Becoming the Product.

Researchers used AI to design 16 working viruses from scratch, harmless ones that only infect bacteria, and the biosecurity experts reviewing the work put it bluntly: the ability to compose genomes with AI now exists, the governance to steer it safely does not.

Swap “viruses” for “any high-stakes system” and the lesson carries straight into enterprise AI. Capability is arriving faster than the rules around it, and the teams that win trust, and contracts will be the ones who build provenance and guardrails in before they’re forced to. Safety is turning into a moat, not a tax. Read the line from the biosecurity team that reframes the whole debate.

💡 The Agent Standards Fight Is Already Being Settled.

The Agent-to-Agent protocol, a shared language that lets AI agents from different vendors work together crossed 150 supporting organizations in its first year, with Google, Microsoft, and AWS wiring it into their platforms and real deployments running in supply chain, finance, and IT ops.

This is the plumbing that decides whether your agents cooperate or get locked inside one vendor’s walled garden.

When the big clouds converge on the same standard, the move is to build for interoperability now instead of betting on a single provider and to watch who’s writing the rules, because they’re shaping your options for years. Here’s who’s quietly becoming the default standard for agents that talk to each other.

💡 The Winning Systems Are the Ones That Snap Together.

This week’s optimists make the case that the frontier isn’t one giant model. It’s modular, composable systems that plug into industry- and government-scale platforms.

For anyone making build-vs-buy calls, the pattern is clear: favor pieces that start as pilots and grow into production without a rebuild, with security and interoperability in from the start. Composable beats monolithic when the ground keeps shifting. Here’s the build philosophy quietly winning right now.

Trending Tools 📈

NationalSecurityAgency/ghidra
(+53 ⭐ per day, 🔷 Java) Link

• Software reverse-engineering framework for analyzing compiled code

• Helps teams: static analysis and scripting for security research, malware triage, and vulnerability discovery

directus/directus
(+16 ⭐ per day, 🔷 TypeScript) Link

• Turns any SQL database into a headless CMS, admin panel, and instant API

• Helps teams: model data and ship internal tools without heavy backend boilerplate

upscayl/upscayl
(+11 ⭐ per day, 🔷 TypeScript) Link

• Free, open-source AI image upscaler for Mac, Windows, and Linux

• Helps teams: sharpen app and doc assets with no vendor lock-in

emilk/egui
(+10 ⭐ per day, 🔷 Rust) Link

• Immediate-mode GUI library that runs on web and native from one codebase

• Helps teams: stand up fast, portable internal tools and lightweight apps

jesseduffield/lazygit
(+7 ⭐ per day, 🔷 Go) Link

• Keyboard-first terminal UI for Git

• Helps teams: cut context-switching and speed up branch and PR workflows

openai/openai-cookbook
(+7 ⭐ per day, 🔷 Jupyter Notebook) Link

• Practical examples and templates for building with the OpenAI API

• Helps teams: jumpstart AI features with shareable, working patterns

invoke-ai/InvokeAI
(+6 ⭐ per day, 🔷 TypeScript) Link

• Creative engine and WebUI for Stable Diffusion, built for production use

• Helps teams: run AI media pipelines with an accessible UI and extensible backend

apache/airflow
(+5 ⭐ per day, 🔷 Python) Link

• Author, schedule, and monitor data workflows as code

• Helps teams: orchestrate complex pipelines with reproducibility and real monitoring

Tech Trend of The Week 📊 

🔍 “Gemini Enterprise” searches jumped more than 5000% vs. last year

Google’s enterprise agent launch is up in the Big Picture, but the search spike shows the other half of it: the people now scrambling to figure out what it is, what it costs, and whether it beats the tools they’re already paying for.

The signal: the gap between “announced” and “understood” is where adoption actually happens. When a launch sends this many people searching the basics, the teams that explain it internally, before procurement does, set the terms.

Our Partner 🎉 

Stop Drowning In AI Information Overload

Your inbox is flooded with newsletters. Your feed is chaos. Somewhere in that noise are the insights that could transform your work—but who has time to find them?

The Deep View solves this. We read everything, analyze what matters, and deliver only the intelligence you need. No duplicate stories, no filler content, no wasted time. Just the essential AI developments that impact your industry, explained clearly and concisely.

Replace hours of scattered reading with five focused minutes. While others scramble to keep up, you'll stay ahead of developments that matter. 600,000+ professionals at top companies have already made this switch.

Join them today, for free.

Hit Reply And Tell Me 💬 

What's the most impressive tech you've seen recently that actually works?

I read every single reply

Go where your customers actually are 📍

Most of the internet copies Reddit.

Blogs, SEO pages, AI models, product reviews. They all pull from the same well.

Reddit is the 3rd biggest site on earth. And it trains the bots your users ask for help.

Odd Angles Media runs Reddit campaigns for over 45 brands each month.

Luckily, we’ve convinced them to give away all of their strategies in a (free) ebook ⬇️⬇️

Grab The Free Ebook (Free) →

Reach the People Who Sign the Checks 💰️ 

40,000+ CTOs and engineering leaders. 96% US-based. 80%+ corporate emails.

They evaluate vendors. They shortlist solutions. They buy.

Advertise with us on Paved → (discount applied)