Claire Woodcock

Just a random thought: A few weeks ago I was in a group of US-based emerging VCs and this person - whom I'm always delighted to have a chat with, asked why there were more minority GPs in the group than the typical and archetype GPs (if you know what I mean). I told him that I didn't have an answer to the question because I wasn't part of the selection process - which was supposed to purely be based on their performance. But in the back of my head was another question: if most of the GPs in the group are of the similar background - e.g., 90%+ of them went to Stanford, Harvard, MIT, etc and they all had a previous career at other VC shops, does the skin color really matter? When they all came from similar paths in their lives, does more skin colors really make the VC ecosystem more diverse? I'm all for having more colors in the VC ecosystem - I'm +1 of that color anyway, but at the same time I just think truly diversifying the VC universe takes more than just adding more colors.

15

9 Comments

Part II of our blog is out! In this installment Abde Tambawala and I explored usage metrics and pricing models across different AI modalities and discussed how traditional User-based subscriptions will need to be supplemented with usage metrics. Take a look and let us know what you think! #pricingstrategy #ai #monetization

27

2 Comments

🤔 How do you make product decisions when you have lots of ideas, some traction, but aren't sure what to build next? Over the past few months, I've spoken to several founders of early-stage climate startups who have all been grappling with this exact challenge. ⚠ They know that focusing on the wrong feature or use case can not only set back their journey toward product-market fit. It could also risk their next round of funding or demotivate their team. Founders at this juncture can get stuck inside the big vision for their business, becoming interested in the many ways they can solve a constellation of problems. If they’re not careful, these leaders will spread their teams too thin by trying five or more different paths simultaneously. How to navigate this? ▶ At Kakuy, our approach is to do one thing at a time, testing and iterating in short cycles until we find the path to product-market fit. 1. [People] Start with the customer. Find a finite group of individuals you’re looking to serve in this next iteration. Narrower is better at the start, because you’re looking for commonalities in their experience that mean many of them will have the same problem to solve. 2. [Problem] Make sure you know what problem you’re solving for your target group. Combine evidence from user interviews, sales calls, product usage analytics, and market research to target the specific problem and what it would mean to your customers if you solved it. 3. [Product] Quickly build a minimum valuable product (or feature) that solves the problem well enough that a customer will pay you something for it. By starting with the customer and quickly validating each of these three aspects of fit, it’s possible to mitigate the risk of wasting time building the wrong thing. 📈 When done right, you can kick off a positive flywheel that generates revenue, which you can then raise against. If you want to find out more, please comment or reach out. #productleadership #productstrategy #MVP

32

4 Comments

if you want to learn about how we built the company and what makes it unique, https://lnkd.in/eQS2VRKf is a really really well written article about us by Bill Kerr - better than what I have managed to put together so far!

34

7 Comments

Nice overview of how we're using data to inform how we invest in early-stage founders. We'll be announcing more this summer but we're seeing some really interesting early data. https://lnkd.in/exU_Juf5

39

3 Comments

Catching up on some of the great presentations from #PEPR24 presentations, and came across Damien Desfontaines’ talk on synthetic data. I obviously disagree with lots of it, but I think Damien generally makes his points in good faith. So I’m going to (try to) do the same in response. The part of his talk that stood out to me as most obviously incorrect and uncharitable was the discussion of the incentives for improved privacy measurements in synthetic data. I’ve included his graphic below as a summary of his argument that there’s no commercial incentive to improve empirical privacy evaluations of synthetic data for context. Sharing three thoughts: 1) If there’s no commercial incentive for better privacy evaluations with synthetic data, differential privacy companies shouldn’t exist. The complexity and form factor of differential privacy imposes enormous incremental friction to configure and use relative to synthetic data. If there’s no commercial case for better synthetic data privacy metrics where none of that friction exists, then how could there possibly be commercial demand for more friction and better privacy metrics — the asserted value of differential privacy? As far as I know, Tumult Labs isn’t a charity, so someone there must believe there’s a market for stronger privacy guarantees — we agree with that, even if Damien doesn’t 😂 2) That’s the theoretical argument, but I’m a synthetic data guy, so I’ll make an empirical argument, too: prospects and customers care deeply about the rigor of privacy tests for synthetic data and it is a buying criteria in *every* deal in our funnel. For *privacy* use cases (more on this in the next paragraph), buyers are not just asking for better metrics; they’re demanding it. My data on this is MUCH better than Damien’s — since he doesn’t actually sell synthetic data to customers — and he’s completely wrong. 3) Despite their website copy, most synthetic data companies are delivering to non-privacy use cases *at least as often* as they’re delivering to privacy-enhancing analysis use cases. Synthetic data is a really useful technology for data quality work, test data management, and anonymization. And in practice, most synthetic data platforms are most heavily used for data quality work (imputing NULLs, class rebalancing, dataset augmentation, etc.). Privacy is — at best — a secondary criteria for these use cases, so the basic privacy metrics are sufficient. At Subsalt, we work exclusively on anonymization use cases and we win deals on the basis of our privacy work. We’re living proof that there’s a business case for more rigorous privacy work in synthetic data. I don’t expect this to solve the eternal (and cursed) debate about empirical vs. theoretical privacy metrics, but I do think it’s worth clarifying that if you want to know about the structure of a market — in this case the market for synthetic data — you should listen to someone that knows at least something about that market.

13

5 Comments

To start, I have to concede that the quote from our website (yes, Subsalt is the anonymous company in Damien Desfontaines’ post; I don’t think he even tried that hard to make it anonymous 😂) overstates the real risks of re-identification in synthetic data. The copy will be changed, because it actually is a good example of sloppy privacy messaging. Mea culpa. But despite being a pretty good dunk, it’s not outcome determinative on the issue we’re actually discussing. The relevant question is: is there market demand for more rigorous privacy metrics in the synthetic data space. The answer is unambiguously “yes.” Customers are not satisfied and they justifiably don’t believe the most commonly used metrics add up to “anonymization.” So why haven’t most synthetic data vendors adjusted their “privacy scores” to accommodate this dissatisfaction? The most commonly referenced companies in synthetic data go to market with freemium versions of their product that can be adopted and used by end-users without a multi-stakeholder purchasing process. The users that actually adopt these tools tend to use them disproportionately for *non-privacy* use cases. And more stringent privacy tests are not neutral to the UX for these users: they’re a source of friction and complexity. So the versions of these synthetic data products you can try off-the-shelf stick to basic privacy metrics that don’t introduce friction. The dirty little secret in synthetic data is that VERY few synthetic data workloads in production are premised on sufficient data anonymization through synthesis. That’s because synthetic data deals where anonymization is the primary use case have traditionally died in the risk, privacy, legal, and compliance departments *because those stakeholders want more rigorous privacy guarantees* that conflict with the needs of the non-privacy users of synthetic data platforms. If every large organization on earth had deployed a synthetic data platform for privacy use cases in spite of this weakness, you’d be able to conclude that there’s a lack of demand for more rigorous privacy work in the space. But all of the synthetic data vendors (combined) have a relatively tiny penetration into the market. The biggest source of today’s “non-consumption” (e.g. the fully uncaptured portion of the addressable market) is the misalignment between the standard synthetic data form-factor and privacy measurements and buyer demand for anonymization. Despite the sloppy copy on our website, that is the latent demand we hear from prospects and customers every day and we take the problem very seriously. I’d be happy to share more about our privacy metrics and approach, because I share the deep concerns and commitment to providing rigorous privacy preserving tools so organizations can do more with sensitive and regulated data — even if I don’t share the skepticism of all empirical privacy metrics.

9

2 Comments

Hearing Jamie Urquhart, co-founder of ARM, speak about their early days was a masterclass in navigating uncertainty and building momentum before product-market fit. A few takeaways stood out: 1️⃣ Vision & Culture First: ARM’s team aligned on a shared vision and trusted themselves to make key strategic calls, even under the pressure of Moore’s Law and fast-evolving tech. 2️⃣ Focus on Burning Needs: They targeted tier 2 customers who could help them establish a foothold, partnering to gain access to markets they couldn’t reach alone. 3️⃣ Ride the Wave: ARM leveraged trends like the shift to 32-bit processors and low-power demand for wireless devices, staying ahead by tailoring their technology to these needs. 4️⃣ Customer-Led Development: ARM only developed features with a committed lead partner. This ensured alignment with market demand and reduced wasted effort. 5️⃣ Grow With Partners: ARM’s disruptive tech turned into business success by leaning on the global reach of their semiconductor partners. The path to PMF is tough, but stories like ARM’s show that clarity, focus, and partnership can unlock massive growth. Thanks to the Deeptech Labs team for hosting this session last week. #DeepTech #ProductMarketFit #Startups

32

1 Comment

Graphy now powers over 100,000 data storytellers 📈 This quote from Stan Massueras really nails it: "Great ideas are only impactful when communicated clearly. Graphy enabled us to transform data into compelling stories, and those stories into meaningful action." Thanks to everyone using Graphy to turn data into stories that drive decisions, inspire teams, and make complex insights easy to act on ❤️

54

10 Comments

// One Average Answer // "Not only could we see a “flattening” and homogenization in terms of what users get to consume (one “average” answer instead of a diversity of results) by virtue of how these systems work; if the input becomes worse, it reinforces the issue. ...all this ultimately provides a pathway for tech companies to further cement their control over information, and potentially enabling them to take over tasks that were once central to, e.g., journalism. Control here to me means: the ability to (1) shape the structural conditions of the public arena (Who gets to see what? Who gets heard? Who gets to organize information?) in the service of economic aims (who gets to benefit?) and connected to political aims (which are often about shaping regulation to entrench and extend one’s control and own position)."

7

4 Comments

Contrarian launch thesis: The best price to launch at (if you took in private funding) is at the last private round valuation. I’ve been thinking about launch mechanics for a while and following Cobie's viral post around fair launches which prompted me to publicly share this thesis. Private value capture has been longstanding debate and as has been identified, you can follow the returns of projects over time had you bought it at the lowest possible public price on the market (see chart below): - ETH: 7,500x - SOL: 300x - OP: 6x - STRK: none Taking higher valuations from private investors is fine (and often very tempting as a founder) but know that you will have numerous overheads, not limited to supply unlocks. You may feel encouraged launch above a threshold price that reflects the time between the private round and public token distribution event. For instance, if the time between the last private round and its public launch is a few years - one would naturally want that price to launch higher. However, these are all artificial/phantom numbers in my opinion. Just as the people behind the project team is its most valuable resource, the community represents its decentralised extension. Thus, if the community is unable to enter at a price that allows them to capture upside, what stickiness does a project have to retain them? With that in mind, I believe community members that put skin in the game should be allowed to participate in as close to terms the latest private valuation. The reason is that many of these participants may have been with the project since Day 1 as a user or contributor. So why should they not be afforded the chance to participate at the same terms as the private investor? In an ideal world, this would resemble an ICO (a true fair launch) but the development of securities laws and how funding rounds are structured (i.e most favoured-nation clauses meaning if a better deal appears after the time of their investment, then that investor is able to default to the preferred terms) makes it impossible to give the community the earliest fundraising round price if the project undertook subsequent ones. Therefore, the next best option is the latest private round to avoid triggering any of the above. A project that launches at a low FDV that allows its community to buy at reasonable prices is going to be far more anti-fragile and resilient when compared against the community member who buys for hype at an inflated price only to experience down-only action. The above thesis is a dynamic canvas representing my three years as a contributor to Drift Labs. There was a time in 2022 when the protocol hit rock-bottom; literally zero. But the community participants who were involved at the projects lowest point, are still involved today and represent its loudest ambassadors. So for god sake, if you’re looking to do a token distribution event, make sure you don’t forget the community that got the project to where it is today.

15

1 Comment

One of the best things about being +7 hours ahead of your founders is being able to plagiarise their analogies before they wake up. Good morning, Stuart Dale 🌞 We're about to bring in several top-tier fractionals to help support growth here at OrbitalX Catching up on recent interviews to give my perspective. I came across this absolutely cracker about startup life and early stage building of SaaS. (see below). 🤣 -------- BTW, if you know of a CMO or fractional leader who is world-class at storytelling and wants a unique opportunity to work across several progressive companies driving performance, a POD P&L, and an amazing culture with everyone in the team winning under your guidance, please get in touch. Founder life! early stage building SaaS. At first, You're the SAS. You're at the frontline dealing with complex problems, often on solo missions to figure stuff out, moving at pace, taking huge bets on yourself and your small team. The risk is high! The reward is high! You then evolve into an army; you know who your commands are, quite a bit more structure and formation. Teamwork and culture need to be strong, but a lot less of a focus is on quick attack, and a more considered mindset is needed. Positive steps consistently win the game rather than heroic moments of greatness. As you grow you evolve into the police. Where you spend more timing policing situations to drive the best outcome. Police get a bad press but the good ones are calm, controlled, pragmatic. Get in touch ------- we're hiring! Mark Choueke James Ford

10

3 Comments

I’m creating a Google Sheet to help super early stage climate founders and I’d love your input. There are many awesome people creating climate solutions. The problem in many cases is they lack visibility and funding. 1️⃣ They’re too busy to make sure they get known by others. 2️⃣ Their projects are too early for them to get enough funding. By creating a database I want to help such founders with: 👀 Exposure I will be making a newsletter issue + a LinkedIn post for 2 of the startups I find the most interesting. 💰 Funding: I’ll share the list with Climate VCs so you can potentially get a good match for funding. Getting early climate projects in front of thousands can help founders in unexpected ways. Being in the radar of VCs, potential employees, partners, etc. Who knows 🙌 If you are an early stage climate founder struggling with visibility and/or funding, I hope this helps you :) To get added to the Google Sheet just fill out this form: https://lnkd.in/eUWPbX4z If you know someone who might benefit from this mention them in the comments or share the post with them.

16

1 Comment

📢 Thrilled to share that Langtrace now has first party support for Inspect AI, an open source evaluations framework that is super flexible to use. What does this mean? Users can now run automated evals on their datasets using techniques like 'LLM as a judge' which are model graded and automatically report the results back to Langtrace for tracking the performance of their applications. In effect, with Langtrace, you can establish a feedback loop from day 1 for continuously monitoring and improving your LLM powered applications. Why is this important? Today, building a product/feature using an LLM is easy to get started and feels magical as it unlocks capabilities that were not possible before. But, the number one challenge developers face is, how to make sure the outputs coming out of their LLM stack is accurate and ensure their users have a good experience. This is why observability coupled with human-in-the-loop annotations and a combination of manual and automated evaluations is key a tooling that you need to have as part of your stack and this has been our number one vision for Langtrace. So how can Langtrace help? Langtrace is open source and it's built on open telemetry standards with auto instrumentation capabilities. This means, developers can get started by integrating Langtrace as part of their stack in just 2 lines of code without spending too much engineering time. Once integrated successfully, Langtrace can - Automatically trace all the LLM, VectorDB & framework level requests. - Groups the LLM requests and presents it to you for manual evaluation and golden dataset curation for assessing the initial performance. - Use the curated datasets to set up and run evaluations on various integration points such as your CI/CD tooling - Github actions, Jenkins etc. - Allows you to version and manage prompts. - Lets you compare and contrast between different models in the prompt playground. This enhances the developer experience of building and scaling LLM powered applications by providing you with rich data and objective metrics in order to gain confidence. Who uses Langtrace? Langtrace is used by various developers, startups and companies ranging from small to large enterprises like Elastic. How do I get started? You have 2 options: - Use the managed and hosted version by signing up on our website. - Self host and run it locally by following the instructions from our github. Links in the first comment below 👇

18

1 Comment

In big companies it’s often not build vs buy. It’s build and buy and do nothing, simultaneously. Many big companies we worked with at Propeller were, at any given moment: • Running internal teams/programs developing competitive products,  • Exploring alternative, overlapping partners, and • Contemplating increasing, pausing or reallocating investment altogether. You’ll do better if you assume big companies aren’t making a binary, or irrevocable, decision.

40

2 Comments

There has to be a fair and transparent process to any change to taxes and a coherent plan for the future. Unfreezing fuel duty would raise £3bn to absolutely dwarf these taxes on ambition and value creation. And benefit the environment.

3

Having co-founded Carbon Re, an AI and ClimateTech spinout of UCL and University of Cambridge, I have first-hand experience of the highs and lows of setting up and running a company in the UK (and trust me, the highs may be high but the lows can be devastating!). If Carbon Re is going achieve our mission to reduce gigatonnes of Industrial CO2 emissions we need a tech ecosystem in which we can thrive and scale fast. I have signed the Startup Manifesto 2024 Open Letter because the Startup Coalition exists to make the UK the best place in the world to start and scale a disruptive company and gives companies like ours a voice. We desperately need disruption to transform our economy and hit our #NetZero goals, so if you haven't signed this already, get involved. See more 👇

45

2 Comments

"As someone who has built platforms most of their life, I am a big fan of the idea that there are no such thing as platforms." Latest episode with ex-CEO of Heroku, Adam Gross Lots of fantastic bits in here. 🔥 Give it a listen

66

12 Comments