Halvar Flake writes the startup guide venture firms will not

Why it matters

Dullien's guide cuts against founder theater: it treats financing, market size, buyer politics and founder stamina as one system, not separate startup lessons.

Thomas Dullien (@halvarflake), the reverse engineer and two-time founder behind zynamics and optimyze, published Halvar's Guide to Entrepreneurship last week, turning two security and infrastructure exits into a manual for software and SaaS B2B founders.

There is a full-circle note here for me as the author: Dullien was one of the figures I first encountered through cracking4newbies on EFNet, the kind of early technical community where reverse-engineering skill traveled person to person long before it was packaged as a startup market. Reading him now as a founder explaining financing, markets and burnout gives the guide an additional edge. It is not just advice from a former operator. It is advice from someone whose technical credibility predates the company story.

The page identifies the guide as version 0.11, published June 23, 2026 and modified June 24. It also offers a PDF version. The framing matters because Dullien is not writing from the usual venture-platform perch. He founded zynamics in 2004, sold it to Google in 2011, spent years inside Google and Project Zero, then co-founded optimyze in 2019 before Elastic announced a deal to acquire it in October 2021.

That resume gives the guide its edge. Dullien is explicit that neither exit was a Silicon Valley-style grand slam. The useful part is that he has seen both paths at founder level: a bootstrapped specialist security tools company and a venture-backed infrastructure company. The guide's central argument is that founders should choose a market and a financing model together, because those two decisions determine the company they will be allowed to build.

A market choice, not a pitch-deck exercise

Dullien divides target markets into four buckets: markets too small to support a company, growing niches, healthy specialized markets, and very large markets. His point is not taxonomy for its own sake. It is that each bucket implies a different capital structure and a different personal outcome for the founder.

In a niche, Dullien writes from experience. zynamics built specialized reverse-engineering products including BinDiff, BinNavi and VxClass, a set of tools TechCrunch described at the time of Google's 2011 acquisition as software for malware analysis, vulnerability research and reverse engineering. Dullien's guide treats that kind of market as potentially valuable but structurally limited: it can produce a durable business, but it is unlikely to produce the return profile that venture funds are built to chase.

The guide is sharper on venture-backed companies. Dullien's napkin math is blunt: if a seed fund is built around a small number of outsized winners, a founder cannot assume that an acquisition returning a modest multiple will make everyone aligned. The VC needs the power-law outcome. The founder, especially a first-time founder with most of their wealth tied up in one company, may rationally care about an outcome the investor considers too small.

That is the guide's most useful warning. The conflict is not that investors are villains or founders are naive. It is that portfolio math and founder math are different. Dullien tells founders to understand that before they sign up for venture funding, not during the board fight that follows.

The VC section is the spine of the document

The funding section is the longest part of the guide, and Dullien says that is because he made mistakes there. He describes fundraising as a herd process in which a company's status can flip quickly from scarce to overfundable and back again. He also warns founders not to confuse investor warmth with commitment: "nothing is real until you have a signed paper in hand."

That line sits inside a broader critique of founder psychology during fundraising. Dullien tells founders to treat verbal commitments as unstable, to be careful with their own verbal commitments, and to avoid using investors as emotional support. The guide even names one section "Investors are neither your friends nor your therapists," which is less a joke than a governance warning.

The lesson tracks with Dullien's own path. Elastic's 2021 announcement for the optimyze acquisition identified him as Optimyze's CEO and co-founder and described the company as an always-on continuous profiling platform for infrastructure, applications and services. Elastic said Optimyze's work used eBPF to provide whole-system continuous profiling with low performance overhead. Dullien's guide now reads that experience back into founder advice: the company story, the market, the financing and the buyer all have to fit together.

Product advice from an engineer who sold to buyers

The guide is not only about investors. Dullien spends substantial space on product, hiring, sales, marketing, co-founder relationships and founder energy. His product advice has the same pattern as the funding section: start with incentives, not features.

He distinguishes technology from product and user personas from buyer personas. A product can delight users and still fail if the buyer has no budget, no political reason to buy, or no internal path to justify the purchase. The example he highlights is FinOps, where a vendor can make buying easier by helping a new internal constituency define itself. The claim is not that category creation is branding. It is that selling B2B software often means creating a reason for an organization to reorganize around the pain your product solves.

That is where Dullien's security background shows. His career was built around tools for specialists, from reverse engineering to fleet profiling. On his about page, he describes starting zynamics around specialized reverse-engineering tools and later moving toward computational efficiency and cloud workload profiling. The new guide extends that pattern into company-building: technical insight is only the starting point, and the founder's job is to translate that insight into a market, a buyer, a team and a financing structure that do not fight each other.

The founder's balance sheet includes energy

The most personal section of the guide is not about equity or product. It is about what Dullien calls a company's three balance sheets: money, technical debt and founder energy. Cash problems can sometimes be solved with investors. Technical debt can sometimes be paid down with hiring and time. Founder burnout is harder to refinance.

Dullien's advice is operational rather than sentimental. Founders need to run at the maximum intensity they can sustain, not at maximum possible intensity until they break. The warning belongs in the same document as the VC math because the pressures are connected. Venture funding buys speed and recruits competition, but it also changes the acceptable range of outcomes and compresses the founder's room to move.

Dullien is careful about scope. He says the guide is based only on software and SaaS B2B, two companies, and no experience operating a company between 15 and 3,000 employees. That caveat is part of why the document lands. It is not a universal theory of startups. It is a founder's field manual from someone who built in narrow technical markets, sold twice, and came back with a warning: before founders optimize for fundraising momentum, they should understand the company that momentum obligates them to build.