“We’re seeing an explosion in creativity and speed”
@rauchg's take on how vibe coding is supercharging the entire playbook for creative builders at #ConnectedStack.
If you’re a builder pushing this frontier forward - moving fast, experimenting aggressively, and surprising yourself with what you can ship - we want to hear from you.
@vercel@trueventures@kevinrose
Most enterprise AI fails for one reason: business teams can't control it.
During #ConnectedStack, @AshwinSreenivas from @DecagonAI shared how they let non-technical users fix AI in 20 minutes.
No engineering. No tickets. No waiting.
This is the unlock that speeds org-wide adoption 👇
"7 days from zero to live AI deployment in large enterprises."
Paul Smith from @AnthropicAI at #ConnectedStack: the average Claude Code implementation time.
Not months of POCs. Seven days.
His take: Domain-specific AI that solves one high-value use case wins. Fast.
This is the moment for focused AI builders.
Humans weren’t built to stare at video feeds all day.
At #ConnectedStack, Filip Kaliszan (@VerkadaHQ) explained how Verkada AI it’s freeing people up to focus on the moments that matter.
“The task of watching video proactively is very mundane... so for us, a very obvious outcome here that we're investing heavily in is the application of both vision language models... and mixing that with large language models to make that information searchable, actionable, and reportable.”
#AI#Security#TrueVentures
As AI agents evolve, one question keeps surfacing: Are we re-engineering our processes to make agents work better or to make humans work better?
Maybe, ultimately, those are the same thing.
At #ConnectedStack, @levie (CEO @Box) said it best to @puneet324 (@TrueVentures):
“Let's actually figure out what agents are able to do, and then let's re-engineer our processes to better support the agent. So it's not that the agent's gonna learn how we work. We're gonna actually be in service of the agent working.”
“You’re able to deliver results quickly on the mechanical side, but it takes more time to capture that mental model from reps — to deliver outputs that are more subjective.”
— @_dan_lee_, CEO and Co-founder of @JoinNooks, in conversation with @kirstenagreen of @ForerunnerVC at #ConnectedStack
The real breakthrough in sales automation isn’t speed — it’s human judgment.
#AI#Sales#TrueVentures
“The battlefield has become the frontend layer.”
At #ConnectedStack, @Vercel’s @rauchg sat down with True’s @kevinrose to discuss how AI innovation is shifting toward the frontend—where open frameworks and flexible architectures allow developers to build with the best models and move faster than ever.
“We’re not pitting humans against AI — it’s one system.”
At #ConnectedStack, @ping_wu of @Cresta broke down how to think about automation in three buckets:
🔹 Tasks neither the business nor customer want — automate them.
🔹 Conversations both sides want but never happen — let AI enable them.
🔹 High-emotion moments — keep them human.
The future of automation isn’t replacement. It’s alignment.
“Agents are tapping into spend that software never could.”
At #ConnectedStack, @puneet324@levie unpack why AI shifts the economics of labor — and why the old “per seat” model doesn’t fit anymore. connectedstack.ai
"As customers start spending in eight figures... usage-based billing really breaks."
@martin_c_mao (@chronosphereio) on why your growing Datadog or Splunk bill is just a symptom of a broken pricing model—and why AI workloads will expose this even faster.
#ConnectedStack#AIWorkloads
The next major breach won't come from stolen passwords.
It'll come from agent identities — and you can't see them yet.
At #ConnectedStack, @RichDandliker from @vezainc laid it out:
Agents are already in your production systems.
They're composite identities — part human intent, part AI autonomy. They don't fit traditional IAM frameworks. And most security teams have zero visibility into them.
The risk? You need to secure BOTH the agent AND the human behind it. Miss either one, your infrastructure is exposed.
Traditional identity security assumes one identity = one human.
That assumption just became your biggest vulnerability.
The companies that figure out agent identity governance in 2025 will survive. The ones that don't will be the cautionary tales.
Moderated by @robby_mtf (@mtf_vc) alongside @VerkadaHQ and @Semgrep — an exceptionally valuable security conversation.
#AgenticSecurity#IdentitySecurity#ZeroTrust#AIAgents
At #ConnectedStack, @AnthropicAI's new CCO Paul Smith joined @joinHandshake CEO @GarrettLord.
Up for discussion:
👉 Why join Anthropic as CCO now?
👉 What does this signal about enterprise adoption scale?
👉 What does it really take to win enterprise AI?
🎥Below ↓
The inaugural #ConnectedStack event brought together the Enterprise AI ecosystem—and showed what’s possible when we build something bigger than any one firm or any one portfolio.
True’s @jcal7 shares why this moment feels like the vintage of a lifetime: trueventures.com/blog/the-vi…