@boldstartvc partnering from Inception with bold technical founders building the autonomous enterprise, weekly newsletter: What's š„ IT/VC šš¼
Joined January 2009
- Tweets 10,469
- Following 3,479
- Followers 20,106
- Likes 33,758
1,477 Photos and videos
š„ up to announce @boldstartvc Fund VII
$250M to back bold technical founders building the autonomous enterprise.
From Inception. Before the world believes.
It always starts with an idea that feels insaneā¦ until it isnāt.
š„šš§µ
104
38
470
86,174
What a Long, Strange Trip Itās Been: SpaceX at $2 Trillion, Software in Limbo, and Venture in 2026
But that uncertainty is exactly where massive opportunities are being built.
This weekās Whatās š„ #502: whatshotit.vc/p/whats-in-entā¦
3
1,719
Going to accelerate even faster with more investment in open source and better hardware - who doesnāt like fast, cheap, secure, private
Jun 8
Narrative violation: according to @Stanford research, local models can answer 71.3% of real-world chat and reasoning queries accurately, up from 23.2% in 2023. Obviously at a fraction of the cost and energy consumption of frontier APIs.
The obvious conclusion: you don't need a frontier model for most tasks. The future is multi-model: local, open-source, smaller and cheaper for the majority of workloads, frontier APIs when no other choices!
2
5
602
well said @peteflorence!
Physical AI is hitting its GPT-2 to GPT-3 moment.
We are at the point where general models that work across tasks will be the ones that take robots from impressive demo to commercially deployed.
This is what @GeneralistAI is building.
Generalist CEO @peteflorence says robotics models are in a transition period similar to the step change between GPT-2 and GPT-3.
They're "starting to cross over into levels of performance where these things are commercially viable for a number of different applications."
"We think this is a crossover point where we have a general model starting to be able to hit levels of reliability, speed, and improvisational intelligence where we can start to get these things out there."
"Very much like ā you take a GPT-2-level model, you scale it to a GPT-3-level model, and certain types of commercial applications start to become viable."
1
4
1,302
What caught my attention this week wasnāt a new model.
It was how quickly the conversation is shifting as AI budgets explode. More enterprises are exploring alternatives: owning context, post-training, routing, and improving open source models.
From Model Wars to Workflow Warsš
This weekās Whatās š„ #501
whatshotit.vc/p/whats-in-entā¦
5
3
9
949
imagine the vulnerabilities and patching needed for all that code š¤Æ
Jun 4
Today, Anthropic engineers on average ship 8x as much code per quarter as they did compared to 2021-2025.
1
1
5
890
huge congrats to the @GeneralistAI team on its $400M raise to build the frontier lab for robot intelligence
the autonomous enterprise is coming for the physical world faster than you can imagine
Amazing what @peteflorence, @andyzengineer, Andrew Barry and team have accomplished in a little over 2 years
we @Boldstartvc are thrilled to continue investing in this new round following our initial inception lead over 2 years ago
Jun 4
We've raised $400M in new funding. This capital goes toward one mission: building general intelligence for the physical world and making it useful to everyone.
1
5
51
5,440
fun to see the post from Partner @ellenchisa laying out why we @Boldstartvc partnered with this team and vision at inception
boldstart.vc/news/generalistā¦
1
3
314
reiterating:
"We're using the more expensive models to explore. Once we scale some of these experiences, we'll look to bring in more efficient models that are more efficient on a token basis or are open source."
wrote more about this here last week: whatshotit.vc/p/whats-in-entā¦
Dara (CEO of Uber) on their AI spend:
"We blew through our AI budget in a quarter, for the whole year. It is forcing us to adjust.
We are going to meter headcount increases because to the extent that my engineers are getting much more efficient, their throughput is increasing. There's a cost to that, and it's a significant cost.
AI adoption has been occurring in all parts of the business āā whether it's engineers and how they scope projects, how they build, debugging, platform migrations.
I'm pushing the teams to fundamentally use the power of AI to rebuild systems and processes from the bottoms up.
I do think it's a combination for us right now of encouraging adoption, but then driving efficiency.
We're using the more expensive models to explore. Once we scale some of these experiences, we'll look to bring in more efficient models that are more efficient on a token basis or are open source."
3
5
22
6,241
Ed Sim retweeted
Jun 1
@edsim Spot on ā continuous scanning at todayās code velocity is simply unsustainable, especially when even Mythos-level spend can burn through $1M in tokens in weeks.
Thatās exactly why last week we launched Evo Continuous Offensive Security (COS) ā AI-native, context-aware pentesting that runs continuously (not once-a-quarter) across AI-generated codebases. Paired with our new Remediation Agent (now in CLI for design partners), these two approaches are complementary: COS finds the exploitable risks in real time; the Remediation Agent reasons over @Snykās security intelligence frontier models to fix them at scale, right in the terminal or PR.
This is what happens when vendors adopt multiple frontier labs, engineer deep security harnesses around them, and bring real context (not just raw model power). As the single-model āMythosā approach cracks under cost and scale, this layered, harness-first strategy is where enterprises will actually get SOTA protection without the budget explosion.
PR blog:
snyk.io/news/snyk-unveils-coā¦
snyk.io/blog/snyk-remediatioā¦
Excited to keep building the harness layer with you and the rest of the ecosystem.
1
1
3
411
Simply unsustainable to do continuous scanning as more code and code bases than ever before
Huge opportunities for security harness engineering to help enterprises get SOTA but cost way less and do more continuous scanning
NEW from me: Anthropicās Mythos isnāt cheapāone tester burned through $1M of tokens in a couple weeksābut companies are still budgeting for it to prevent hacks: theinformation.com/articles/ā¦
9
8
1,267
Whatās š„ issue #500 š¤Æ
Honestly hard to believe.
What started as notes to myself shared with a few friends somehow turned into 500 straight weeks of writing and reflecting on enterprise tech, cybersecurity, AI, infrastructure, and venture. That's over 9 1/2 years!
šš¼ to everyone who reads, shares, replies, and keeps me learning every week.
This issue: why enterprise AI is entering āPhase 2ā as the subsidy era ends and enterprises shift toward ROI and outcomes.
Still very early š„
full post next...
6
2
20
3,647
full post here: The End of the AI Subsidy Era and the Rise of ROI Per Workflow
whatshot.substack.com/p/whatā¦
2
2
2
731
agreed - overblown but...
reality is we're entering the next phase where enterprises fully aware of costs as vendors charge more for consumption and subsidy era is over
We'll see much more in way of intelligent routing to models, more investment in open source for hybrid workflows - SOTA as brain, offload to OS where makes sense, and how to measure what success is
still early and more to build
May 29
I'm suspicious of that that whole story about Uber blowing their AI budget and being disappointed in the results - I dug into it and it appears to have been built on very shaky foundations
ALT I started digging into this in response to a growing volume of stories claiming that large companies were sounding the alarm because their AI usage costs had grown so large. The most widely cited of these stories appear quite overblown to me. The most discussed has been Uber, based on this report where CTO Praveen Neppalli Naga indicated that Uber had āmaxed out its full year AI budget just a few months into 2026ā, mostly thanks to Claude Code. Given that Claude Code only got really good in November itās entirely unsurprising to me that a budget set in 2025 may have failed to predict demand for that tool in 2026! That Uber story was further fueled by comments made by Uberās COO, Andrew Macdonald, on the Rapid Response podcast. I tracked down the segment and there really isnāt much there. Hereās what Andrew said:
ALT But then you sometimes go and talk to your senior engineering leaders and youāre saying, OK, how many projects that were on the cutting room floor got moved above the line because of the productivity gains because 25% of our code commits were via Claude Code last quarter? That link is not there yet, right? I think maybe implicitly thereās more that is getting shipped. But itās very hard to draw a line between one of those stats and, OK, now weāre actually producing like 25% more useful consumer features, right? And that line is hard to draw. Somehow this fragment turned into headlines like Uberās COO says itās getting harder to justify the money spent on AI tokenmaxxing, because the market for stories about AI failures remains enormous.
5
5
19
3,976
Enjoyed my conversation with @MariaDerchi on the @MiamiTechPod
Talked about what we look for in founders, our 5 Ps for inception investing, AI investing and building, why everyone doesn't need to chase VC š°, and why it's early but promising in Miami and parallels to NYC tech ecosystem back in the day (mid-90s š)
Listen in!
May 26
What makes someone fundable?
Top seed investor @edsim of @Boldstartvc breaks it down.
Including:
ā why most founders shouldnāt raise VC
ā founder green and red flags
ā what AI changes about startup building
ā why Miami is just getting started
š§ New Episode: miamitechpod.com/what-top-seā¦
1
2
9
1,035
Whatās š„ in Enterprise IT/VC #499 is live!
The More We Automate, the More Valuable Humans Become
read more on why @danshipper scaled his team from 4 to 30 with more agents and why Cloudflare and @eastdakota is rearchitecting his enterprise and hiring more humans to build and sell
must read as everyone strives to agentify their orgs at rapid pace - more humans needed as more work and more decisions to be made - counter to prevailing narrative
also reminds me of this - "a computer can never be held accountable"
3
3
9
5,484
full post here: The More We Automate, the More Valuable Humans Become
whatshotit.vc/p/whats-in-entā¦
1
362
The surface area is only expanding and more humans needed to patch and fix...
which is why dozens of security companies have partnered with Anthropic - the market opportunity for the right cybersecurity companies is only getting bigger and more urgent, not less
May 22
Last month we launched Project Glasswing, our collaborative AI cybersecurity initiative. Since then, we and our partners have found more than ten thousand high- or critical-severity vulnerabilities in essential software.
1
4
1,615
more written here a couple months ago x.com/edsim/status/203787845ā¦
1
473
must read as everyone strives to agentify their orgs at rapid pace - more humans needed as more work and more decisions to be made - counter to prevailing narrative
also reminds me of this - "a computer can never be held accountable"
May 21
Weāve automated every single thing we can @every with AI agents.
And yet thereās way more human work to do than ever. Weāve gone from 4 -> 30 human employees since GPT-3.
I wrote a report on the structural reasons: how AI makes expert competence cheap, why that drives up demand for experts, and why the dynamic only intensifies as we approach AGI.
After Automation: every.to/p/after-automation
4
1
26
11,065