Running to stay in place

Two weeks ago, I was at a dinner with a bunch of growth and go-to-market leaders. At one point, somebody posed the question: if your CEO gave you additional budget and complete freedom to spend it, where would the next dollar go?

The consensus choice was more events. Not just conferences, though those can be good. Even throwing a small, low-stakes event in your city (matchas and a roundtable conversation, pizza and a fireside chat) is worth the effort. People are just desperate for in-person connection.

Some people mentioned cold calling. (The trick: few people will convert on that call, but if you track them, they’ll show up on your site a week or two later and might just convert direct.) A few others mentioned influencer marketing. Everyone said that cold email is dead and paid is too expensive.

The meta-point everybody made again and again is that a lot of what used to make marketing tactics effective depended on them being difficult to do. And the problem is that AI is making a lot of those things too easy.

Here’s an example: cold email never worked great, but it used to work when you put in the effort to deeply research the prospect. Whether consciously or not, the recipient recognized the investment, which caused them to stop and give the email an additional second of their attention, and that was your opportunity. 

Now AI can do all of that work in seconds. The output might be just as good (honestly, probably better), but it’s completely and utterly undifferentiated. Everybody’s inboxes are full of equally personalized messages. As a result, that personalization tells the recipient much less about whether you actually care or understand their problem.

This whole conversation reminded me of trying to scale paid marketing at Airtable back in 2018. The advice from people who were exceptionally good at paid was surprisingly consistent: trust the algo. What that meant was you didn’t win by trying to find your exact audience and targeting them specifically. You won by giving the algorithm better inputs, namely, a strong audience to build lookalikes from, and lots of creatives to test.

Realizing this, the best teams out there had built extraordinary production pipelines, generating hundreds or thousands of creative variations every week. At the time, this was a serious technical and operational achievement. It required significant investment of both people and systems/tooling. These were engineering teams masquerading as paid marketing teams.

Today, “generate a thousand ad variations” is a task I can hand off to my agent during breakfast and review when I sit down at my desk. Even building the pipeline (i.e. connecting your internal systems to the ad platforms) is not particularly hard anymore.

That doesn’t mean creative testing doesn’t work any more. But the ability to produce a significant volume is now table stakes. You will now need to differentiate by understanding the customer better, developing a more compelling concept, or iterating more quickly and smartly through variations.

If you haven’t had a chance yet, go look at WisprFlow’s (now viral) influencer marketing playbook. There are a lot of parallels to the evolution of paid marketing. The winners of early influencer marketing figured out how to generate great scale by managing large programs across thousands of influencers. WisprFlow’s current winning playbook is not just about scale, but insane levels of quality at scale. That’s still hard. And that’s why it works.

The uncomfortable implication for all of us is that AI adoption is probably not going to buy any of us days of leisure. Automating a process can make you or your company much more productive. But if everyone else can automate the same process, that productivity gain may do little to improve your competitive position. You’ve got to run just to stay in place.

This is particularly clear in the world of growth, because it is so competitive and any tactical advantage lasts months at most. But this will be true across the entire organization. 

So the job is to keep asking what remains difficult and valuable as the tools improve. Difficulty alone isn’t a reason to do something. There’s plenty of expensive, manual work that accomplishes little. But when you find something your customers value that competitors struggle to do, there may be an opening.

If you can figure out how to scale it before everyone else does, you might have a few months advantage. Then you have to go do it again. That was the opportunity with superscale creative pipelines back in 2018. The question is where the equivalent opportunity sits today.

David Peterson

Could the future of software be fluid
How do we get the best of AI without losing the soul of software?

The future belongs to young missionary teams
Why it makes more sense betting on youth in the current moment

The AI-native enterprise playbook
Ten real-time observations on a rapidly evolving playing field

No more painting by numbers
It’s the end of the “SaaS playbook.

HARD TECH

AI Infrastructure Debt Concerns Grow Among Investors. Wall Street is growing increasingly wary of the debt financing fueling the AI infrastructure boom, questioning if massive spending will yield sufficient returns. This concern is particularly acute for 'neoclouds' and data center builders, evidenced by CoreWeave's 2032 bonds seeing yields surge to nearly 15%, causing its stock to tumble 17%. The Bank of England Governor warned markets risk a correction by overestimating AI returns, signaling potential fragility in the AI investment landscape.Axios also reported that “Gartner predicts that more than 80% of specialized AI cloud providers, known as neoclouds, will cease operating independently by 2030.”

Google launched its new Argon model, and many view it as placing the company solidly back on the frontier. That said, Michael Parekh wrote that we should “expect frontier leadership to keep ebbing and flowing, in an infinite game, not a race. Google is unique, with a full vertical stack from chips to billions of users, so it can afford to arrive late, and cheap. Its core goal is still ‘good enough’ models for the billions”

Workday has agreed to acquire C1.ai, formerly ConductorOne, an enterprise software company focused on managing human and AI agent identities, which previously raised $111 million. This acquisition enables Workday to enhance its offerings by addressing the complex identity management challenges brought on by the increasing use of AI agents in enterprise environments. The deal signals Big Tech's strategic focus on integrating AI governance capabilities. “Workday presents the acquisition as a natural extension of its existing role. Hiring, job changes and departures are already recorded as Workday events, giving the platform timely signals for adjusting access. By adding C1.ai, Workday intends to become what it calls “the identity governance layer for people, money, and agents”.”

Intuit Acquires UK Fintech Sequence for $200M. Intuit has acquired UK-based fintech Sequence, a startup using AI to help companies manage revenue, in a deal reportedly worth $200 million. Sequence had raised $19 million in 2022, led by Andreessen Horowitz. This acquisition enhances Intuit's financial software capabilities and marks one of the largest exits for a UK AI startup with a female co-founder.

HARD MARKETS

Blue Origin Commits to New Glenn Rocket Scale-Up. Blue Origin is 'going all in' on its New Glenn reusable heavy-lift rocket, with infrastructure designed to support multiple launches per week by the end of the decade. This aggressive production target signals a major commitment to increasing launch cadence and market share in the commercial space sector. The move intensifies competition with rivals like SpaceX, indicating a robust expansion in private space capabilities. “Blue Origin is expanding its rocket facilities, while SpaceX is pulling back from the commercial launch market with its Falcon 9 and Falcon Heavy vehicles. As Ars has reported, a growing launch crisis is emerging as commercial satellite customers have struggled to purchase a Falcon launch after 2028. This has created an opening for medium- and heavy-lift vehicles that, if Blue Origin can get New Glenn flying successfully and frequently, it is poised to fill. The company’s previous investments, along with the new infrastructure, suggest Blue Origin founder Jeff Bezos is deadly serious about making this a reality.”

US, Taiwan Ink 13 Drone Tech Supply Chain Deals. US defense companies and startups have signed at least 13 deals with Taiwanese hardware firms since September 2025 to build a secure drone technology supply chain. This collaboration aims to create uncrewed boats and aircraft using a 'non-red' supply chain. The initiative strengthens Taiwan's defense capabilities and reduces reliance on components from adversarial nations. “The emerging industry has the potential to serve the U.S. military’s demand for a secure drone supply chain that steers clear of China, while providing Taiwan with an arsenal that it sees as essential to its survival as a self-ruled democracy.”

WindBorne Wins $50M US Air Force Contract for Balloon Data. Silicon Valley startup WindBorne secured a nearly $50 million contract from the U.S. Air Force for its weather balloons and AI-powered weather prediction technology. This substantial defense contract highlights the military's increasing reliance on advanced AI and meteorological data for strategic operations. It also showcases the dual-use potential of cutting-edge climate and atmospheric monitoring solutions for national security. “WindBorne currently has about 700 weather balloons flying around the globe, collecting atmospheric sensor readings that the company sells to customers and also feeds into its own AI weather forecasting model. WindBorne, founded in 2015 and based in Redwood City, California, is one of the first organizations to directly feed empirical data into a deep-learning weather model.”

HOW TO STARTUP

UK to Ban Non-Compete Clauses for Employees. The UK government plans to curb non-compete rules, eliminating clauses that typically force employees to wait before moving jobs or starting new companies. This policy change is expected to significantly benefit the UK startup sector by fostering innovation and increasing talent mobility. “Dom Hallas, Startup Coalition CEO, said: “We think this is a huge move and so do lots of friends including so many of the great founders that signed the letter with Inherent just the other week. Because talent matters and mobile talent means startups can build. The Valley was built on free-flowing talent. In the UK, we’re now seeing our own versions of this story with companies being built from alumni of both homegrown firms like Revolut and beyond. But we need that to move faster. Hiring and building faster means British startups and scaleups can grow faster. Barriers to that slow our ecosystem down.””

HOW TO VENTURE

The Death of Venture Capital’s Short Kings. In an important piece, Kyle Harrison argues that while bigger funds can support larger ceilings, they don’t create proportionally more winners, reinforcing a system that prioritizes “building what’s fundable” over durable businesses and leaving the long tail increasingly underserved. Benchmark and USV—the last major “short kings” or “Cottage Keepers” of venture that deliberately stayed small—have both scaled up, with USV raising $900M (its largest cycle) and Benchmark raising $2B including its first growth fund, marking the end of the pure small-fund model amid capital concentration and AI-driven trillion-dollar outcomes. “The unfortunate reality is that rather than embrace alternative ways of building businesses (permanent capital, HoldCos, revenue-based financing, or just bootstrapping a calm company for longer), we’re going to keep reinforcing the normative funnel of “building what’s fundable.” Startups of a certain shape and shimmer, built to look like what the next round wants rather than what customers need, necessarily. And I think that’s net negative for the broader business building mechanism. Now, the more relentlessly capital concentrates at the top, the larger and more underserved the long tail at the bottom becomes. Somebody is going to figure out how to serve it. I hope they do. But they’ll have to do it with two less short kings.”

CruxOCM
Advanced Process Control Engineer (Vancouver, Remote)

FalkorDB
Core Developer (Tel Aviv)

Reco
Product Manager (Tel Aviv)