💡 The $36M Bet That Someone Will Try to Phish You Tomorrow

In partnership with

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

Featured Tool 📊 

Stop letting your bank account earn almost nothing.

You have money sitting in your bank account that could be earning. But moving it around is a hassle. Rivo automatically moves the cash you are not using into U.S. government-backed assets and brings it back before your bills are due. Set it up once, and let Rivo do the work. Try Rivo for free now.

Sign up for free →

Quick Hits 🎯

• 🍃 IBM had a rough quarter, then went on record insisting AI isn't killing the mainframe. The truth is messier than either headline

• 🎙️ OpenAI just gave Codex and ChatGPT full-duplex voice control, so you can talk to your codebase instead of typing at it

• 💼 ServiceNow dropped $40M on an Indian banking software specialist, a small check with a big signal about where fintech AI is headed

• 🖥️ Microsoft had to publicly respond after LG monitors started installing McAfee ads directly onto Windows

• 🎭 Meta's newest AI ad pairs an apocalyptic Bowie track with an optimistic pitch and the internet noticed the whiplash

• 🔓 A security incident tied to OpenAI and Hugging Face raised fresh questions about provenance and trust in open model ecosystems

• 📉 Monday.com laid off hundreds of employees to go all-in on AI-native features

🎁 + 5 other stories you might find useful

Our Partner 🎉 

Win AI Search Without a Big Team

92% of VCs use AI to find companies. 58% of buyers start there, too. If you're not showing up in AI answers, you're invisible before the conversation even starts. Join HubSpot for Startups, Anthropic, and Marketing Against the Grain on July 30 (11 am ET) for a live AEO teardown. Real startup. Real recs. Register and unlock the free Startup Visibility Bundle.

Register Now

The Big Picture 🖼️

💡 The Mainframe Just Bought Itself More Time.

IBM's quarter looked rough on paper, and the market assumed AI workloads were finally finishing off the mainframe for good. Instead, IBM pushed back hard, framing the shift as a budget reshuffle rather than a retirement.

What's actually happening is more interesting: enterprises are running AI and legacy systems side by side, and the winners will be whoever makes that pairing least painful. You can read where IBM is placing its bets in after shocking quarter, IBM insists AI isn't killing the mainframe.

The takeaway: legacy infrastructure isn't going away just because AI got interesting. It's becoming the thing AI has to work around.

💡 Voice Just Became the Next IDE Battleground.

OpenAI shipped full-duplex voice control for Codex and ChatGPT on desktop, meaning your coding assistant can now listen and reason in real time instead of waiting for a prompt.

This isn't a novelty feature, it changes the shape of pair programming, code review, and debugging sessions when the assistant is a live participant instead of a chat window.

The engineers who get comfortable talking through problems out loud will move faster than the ones still typing full paragraphs. See what's different in how GPT-Live's voice control reshapes agentic coding.

The takeaway: the interface war in developer tools just expanded past the keyboard.

💡 Platforms Are Buying Trust, Not Just Features.

ServiceNow's $40M bet on an Indian banking software specialist isn't really about the technology, it's about credibility and speed in a market where local regulatory knowledge matters more than raw AI capability.

Watch for why platform players are choosing regional partners over building in-house, because this playbook is about to repeat across every regulated industry AI touches.

The takeaway: in fintech AI, trust is now a build-versus-buy decision, and buy is winning.

💡 OEM Bloatware Just Became a Trust Problem, Not Just an Annoyance.

Microsoft had to publicly respond after LG monitors shipped with McAfee ads baked into Windows.

That's not a UX complaint anymore, it's a signal that platform owners can't fully control what happens once hardware partners get their hands on the OS.

If you're evaluating vendor relationships in your own stack, this is worth reading before your next procurement cycle.

The takeaway: platform trust is a maintained asset, not a default setting.

💡 Emotional Marketing Is Doing the Work Product Specs Used to Do.

Meta's newest AI ad leans entirely on an apocalyptic Bowie track and an optimistic message, with almost no actual product detail.

It's a bet that feeling matters more than features right now, and it's a bet that invites scrutiny about tone, safety, and whether the emotion matches the reality. The full ad breakdown is worth ten minutes if you're thinking about how your own company talks about AI.

The takeaway: in a crowded AI market, brand feeling is becoming its own competitive lever.

💡 Open Ecosystems Are Only as Strong as Their Weakest Handoff.

A security incident connecting OpenAI and Hugging Face is a reminder that as models, weights, and tooling move across more hands, the surface area for something to go wrong grows with it.

Here's what actually happened and why alignment conversations keep circling back to paperclips, a useful read whether you ship models or just consume them.

The takeaway: interoperability and governance are becoming the same conversation.

💡 Spear Phishing Just Got an AI-Native Opponent.

AegisAI, founded by former Google security leads, raised $36M specifically to catch AI-generated phishing before it lands.

As attackers use AI to write more convincing, more targeted messages, defenders are having to match that speed with anomaly detection that doesn't drown teams in false alarms. The founding story and what they're targeting first is a good read if security is anywhere on your roadmap this quarter.

The takeaway: AI didn't just create new attacks, it created a new category of defense.

💡 Layoffs Are Becoming a Strategy Statement, Not Just a Cost Cut.

Monday.com cut hundreds of roles specifically to redirect resources into AI-native features, trading breadth for depth.

It's the same move several mid-size SaaS players are making right now: fewer people, sharper focus, faster AI shipping. The full breakdown of what they're prioritizing instead tells you where the rest of the category is likely headed next.

The takeaway: in SaaS right now, "AI-first" increasingly means "smaller, on purpose."

Trending Tools 📈

skypilot-org/skypilot (+11 ⭐ per day, 🐍 Python) Link

• A single platform for running AI workloads across every cloud and on-prem cluster you've got

• Helps teams stop babysitting fragmented compute setups and start treating GPU capacity like one shared pool

aquasecurity/trivy (+20 ⭐ per day, 🦙 Go) Link

• A vulnerability scanner that covers containers, Kubernetes, code repos, and cloud configs in one tool

• Helps teams fold security scanning directly into CI/CD instead of bolting it on after the fact

ggml-org/whisper.cpp (+70 ⭐ per day, 🗣️ C++) Link

• A C/C++ port of OpenAI's Whisper that runs fast, offline speech recognition on local hardware

• Helps teams ship voice features without sending audio to a third-party API

anchore/syft (+9 ⭐ per day, 🔎 Go) Link

• Generates software bills of materials straight from container images and filesystems

• Helps teams answer "what's actually in this build" when a vendor or auditor asks

DioxusLabs/dioxus (+347 ⭐ per day, 🦀 Rust) Link

• A fullstack app framework for web, desktop, and mobile from a single Rust codebase

• Helps teams cut the number of frontend stacks they maintain without giving up performance

jj-vcs/jj (+45 ⭐ per day, 🦊 Rust) Link

• A Git-compatible version control system built around a simpler, less error-prone workflow

• Helps teams onboard engineers faster by removing Git's sharpest edges

schollz/croc (+898 ⭐ per day, 🚀 Go) Link

• Securely copies files between machines with almost no setup

• Helps teams move large datasets or build artifacts between environments without spinning up infrastructure for it

feast-dev/feast (+10 ⭐ per day, 🧠 Python) Link

• An open-source feature store that manages and serves ML features consistently

• Helps teams keep training and serving in sync so models don't degrade quietly in production

Tech Trend of The Week 📊 

🔍 "Codex" spiked this week

The jump tracks almost exactly with OpenAI's full-duplex voice rollout for Codex and ChatGPT desktop. Developers who'd mostly ignored Codex as a side feature are suddenly searching for what changed and whether it's worth switching workflows over.

The signal: a single feature update was enough to pull an existing product back into the search conversation. Teams that figure out what actually improved versus what's just getting attention will make faster calls on whether to adopt it.

Our Partner 🎉 

While your trucks are running, calls are going to voicemail.

Every missed call is a job your competitor just booked. Podium's AI Employee responds in under 2 minutes, qualifies the lead, and schedules the job — while your crew keeps working.

See It In Action

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)