Proactive board management for year-end
Gil Dibner

With six weeks left in the year, what’s the best way for an early-stage founder to manage their board at this crucial time of year? There’s no one-size-fits-all answer, but here are a few concrete ideas:

Updates. If you haven’t started sending regular (weekly or monthly) updates to your investors and other stakeholders, you should do this immediately.  In my experience, it’s highly correlated with company performance. It also buys you a lot of credit with your investors, helps you avoid surprises, and ensures everyone is on the same page when they do meet.

Do you need a board meeting right now? If you have a board meeting set between now and the end of the year, think hard about whether that makes sense. If things are humming along, it’s probably fine. If that board meeting is not set up correctly, however, it might be more trouble than it’s worth. A board in December falls at a time when operators (like you!) should be focused on closing their quarter. Investors are often distracted by other portfolio companies or holiday vacation plans. More importantly, you probably won’t have final results for December, so catching up in January might make more sense. Decide on your desired cadence for board communication for the coming year, and work with your board to set those meetings up for the next 12-18 months. Use the schedule and cadence of board meetings and updates to best serve you as you manage the business.

Decide (for yourself!) what truly matters. Unless everything is going up and to the right (in which case, congratulations!), most companies at the early stage are struggling with an ever-shifting set of problems and challenges. This can be challenging for boards to navigate, as there is always a tendency to be reactive. I’ve seen countless board meetings go off the rails when an investor sees things not going well and tries to encourage a founding team to explore various (random) options to fix it. This can be massively distracting and confusing. The key unlock for founders (particularly first-time early-stage founders) is to take control of the board dynamic.

Be your own toughest board member. Rather than wait for a board meeting to take place and roll the dice on how that conversation unfolds, work to shape that conversation yourself. It’s November now. Imagine the next four board meetings starting January 2026. What do you currently believe are the most crucial things the company needs to achieve by the time you have each of those meetings? These could be specific sales targets, customer milestones, product/engineering achievements, key hires, or anything else. Work backwards to today. What are the key KPIs or OKRs against which you can manage right now? If, for example, there is a goal of “hiring a product leader by 2Q26,” how can you translate that into an aggressive plan of action today? How would execution against that plan look as reported in your new weekly update (see above)?

Throw out old targets. For any founder who hasn’t had a lot of board/investor interaction, there tends to be a lot of confusion over targets and forecasts. A target is what you realistically want to achieve. A forecast is what you currently believe you will achieve based on the facts and data you currently have. As you approach the end of the year (or the quarter), it's crucial to avoid the temptation to inflate forecasts. It’s also crucial to avoid clinging to really old targets that are no longer relevant. Pulling out a target you set 12 months ago and measuring yourself against that today can be informative (and is often sobering), but it can also set up a conversation to be needlessly painful and unhelpful. Forecasts should be accurate and targets should be realistic. If not, the numbers quickly become meaningless.

Get a head start on accountability. Once you’ve decided what matters, hold yourself to account proactively so your board doesn’t feel a need to take on that role. Working backwards from your 1Q26 goals, what has to happen in November and December of this year to ensure you are on target? Start holding yourself to account against those short-term objectives. Even if it’s too late for you to drive up your year-end revenue numbers, you might be able to take other steps that set you on the right path to hit your 1Q goals. This also allows you to control the narrative and set the tone as you approach year-end. Strong updates in November and December that show you are executing against your interim objectives could set you up optimally for constructive board conversations in 1Q. Create some early wins. They will compound.

Give them homework. As you approach year-end, send out your draft budget and plan for the coming year, along with your specific requests for assistance (hiring, advice, etc.) and the key strategic topics you plan to address.

A useful forcing function. At the end of the day, a board is just another tool a founding team can use to help them effectively run their business. Early-stage startup boards can easily become noisy distractions, but they can also serve as a crucial team of trusted advisors that convenes as a forcing function for accountability and strategic clarity. As a CEO, the best way to ensure that happens is to take responsibility. Proactively design the board meeting schedule, interim reporting cadence, KPIs, and agenda to help you achieve the best results you can achieve. Instead of scrambling to prepare for another year-end board meeting, consider that next month might be better spent driving results in your business and carefully re-evaluating how you are using your board as a strategic and tactical tool.

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.

The Age of Artisanal Software May Finally be Over
Every wave of technological innovation has been catalyzed by the cost of something expensive trending to zero. Now that’s happening to software.

Founders as Experiment Designers
David on why founders should run everything as an experiment.

ENTERPRISE/TECH NEWS

Satya’s AI bet. Dwarkesh and Dylan use a tour of Microsoft’s new Fairwater 2 “AI factory” with Satya Nadella to unpack how hyperscalers are really thinking: 10x training capacity every ~18–24 months, but with fungible, multi-model infrastructure rather than being a single lab’s host. Nadella frames AI as a long, messy, industrial buildout where value may shift from pure models to agent “scaffolding,” data, and observability, and where economics become “per user + per agent.” For founders and VCs, this is a blueprint for where margins may actually accrue: agent control planes, sovereign/vertical clouds, and infra that helps enterprises manage fleets of autonomous coworkers, not just call an API.

Bezos back in the game. Bezos is jumping back into day-to-day company building as co-CEO of Project Prometheus, a newly formed AI startup with a staggering $6.2B in funding and ~100 people already onboarded from OpenAI, DeepMind, and Meta. The company’s focus is “AI for the physical economy”—using advanced models to simulate and optimize real-world engineering and manufacturing across sectors like aerospace, autos, and computing, in the same vein as outfits like Periodic Labs that are building AI-first scientific labs.

Zuck’s Biohub. In an interview with Latent Space, Priscilla Chan and Mark Zuckerberg explain how CZI’s Biohub is evolving into a “frontier AI x Bio lab,” using a 10,000-GPU cluster, EvoScale’s ESM-3 team, and massive open datasets (like the Human Cell Atlas) to build virtual models of cells and, next, the immune system. Their bet is that turning biology into an in-silico, modelable system (via new microscopes, imaging pipelines, and AI models like VariantFormer, CryoLens, scLDM, GREmLN, and rBio) will radically speed up basic science. CZI itself won’t build consumer therapies, but aims to arm scientists and partners with the tools to enable true precision medicine and, over time, “cure and prevent all disease.”

HOW TO STARTUP

Cursor dispatch. Cursor has rocketed to a $29.3B valuation on the back of its AI coding environment and in-house Composer model, making it one of the purest examples of the current “capital as weapon” race in AI dev tools. Brie Wolfson is embedding with the company for sixty days to document how an AI-native, model-dependent product org actually operates from the inside. Together, the news and the dispatch set up Cursor as a live test case for whether owning the IDE, the model, and the workflow can justify decacorn economics.

“Dumbest idea I’ve heard.” Grant Lee (CEO of Gamma) tells the story of building what every AI founder says they want: a wildly fast-growing, profitable AI startup with a tiny team, in a “boring” category most investors dismissed. He goes deep on how they actually got there, like by ruthlessly rebuilding onboarding until word of mouth exploded, doing founder-led influencer marketing with thousands of micro-creators, stress-testing every idea with real users in hours, and running 20+ models under the hood to make a “GPT wrapper” that’s defensible. It’s also one of the most concrete playbooks I’ve seen on org design in the AI era: lean, generalist “player-coaches,” painfully slow hiring, and an almost pathological focus on taste, brand, and unit economics from day one.

HOW TO VENTURE

China’s rise. Martin Casado was quoted in the Economist saying that there’s an 80% chance, when an entrepreneur walks into their office at a16z, that they’re building on top of Chinese open-source models like DeepSeek. Why? Dave Friedman argues that the reasons are to escape platform risk, U.S.-driven governance constraints, and rising API costs. Friedman frames model choice less as “which frontier model is best?” and more as a strategic decision about owning weights, optimizing inference, and avoiding a future where your product can be sherlocked or shut down by U.S. vendors or regulators.

Build what’s fundable. Kyle Harrison argues that Y Combinator has drifted from its original mission of helping outsiders navigate an opaque tech world into a Consensus Capital Machine optimized for manufacturing fundable companies rather than backing important, contrarian ones. He shows how the Request for Startups, founder demographics, and recent “slop startup” controversies reflect a broader shift in tech toward hyperlegibility, status-seeking, and building what looks good to VCs instead of pursuing meaningful quests. 

DualBird raises $25M to rebuild the foundation of analytics and AI data infra.

Reco recognized as a CRN 2025 Stellar Startup.

Candu
Full Stack Developer (Remote or London)

DualBird
DevOps Engineer (Tel Aviv / Haifa)

Paradime
Chief of Staff (NYC, London)