I was catching up with a few institutional LPs (those who invest big dollars into VC funds) last week and the sentiment at the moment from many investors is cautious optimism. Yes, this is one of the biggest platform shifts in history but yes, some of these prices are outrageous, and many companies who are early leaders now can just as easily be vaporized with the next update. Add to this the incredible pace of building, exaggerated claims of what products can offer, and one can easily get whiplash from all of this noise. To that end, here’s a reminder for all of us -

Naveen, VP GenAI at Databricks, summarizes the week that was.

More from Jim Fan at Nvidia…

Isn’t it a wonderful time to be a founder? Just make sure your architecture easily allows you to swap out a model, and if you don’t like the models or the cost, just wait a minute, it will get better and cheaper.

Speaking of hype, many a startup building AI into their developer offerings were certainly alarmed by the claims of Devin a few weeks ago, but it turns out that much of it is well…not exactly true. I applaud moonshots but seems like this company stretched a bit on its ability to deliver.

Another reminder for founders and investors, AI is a technology and not a product. Solve a problem 10x better than what already exists for a user and buyer willing to pay and make sure that they love the product so much that you can’t pry it out of their cold, dead hands for real PMF.

And one last thing as a founder or investor, especially at Inception, make sure you don’t over rotate to Total Addressable Market Size or TAM - you can miss lots of opportunities.

Tomasz shares some similar lessons on TAM from Tobi, founder and CEO of Shopify…

Now zoom in, build that product…

A wise founder also told me, especially in new markets, can you take your first 10 customers, and double each of their revenues in next year or two? if so, you got a market

It’s different, by the way, depending on new category creation or simply going 10x+ for incumbents. This from What’s 🔥 #277:

With that, it got me thinking about two types of founders we like to partner with @boldstart, category creators which represent the bulk of what we do and ones who reimagine existing markets like Kustomer. Here’s a quick summary of the pluses and minuses of each type and how we think about day one investing in those.

Category Creators

1. Market timing risk

2. Market size unknown - need to imagine what it can be

3. Customer education on market required

4. Less competition, usually replacing home-grown solutions

5. Easier to get initial product out the door since your startup is creating the base features required

6. When market clicks, it can scale rapidly

Reimagining Existing Markets

1. Market already exits

2. Market size known and usually huge

3. Customer education on why your specific product required

4. Tons of competition - need to differentiate, characterized by RFPs and competitive bake offs

5. Feature parity required before adding extra bells and whistles - costs 💰

As always, 🙏🏼 for reading and please share with friends and colleagues.

Scaling Startups

#more killer advice from Jensen from 2003! Play long ball or as I like to say, sell the product (today), market the vision (the future) - must watch video

“I don't need to change the world overnight, I am gonna change the world over the next 50 years”—Jensen, NVIDIA CEO, 2003

#what Sundays are for - Jamie Dimon, CEO JPM Chase

One exercise that I find useful (and sometimes painful) is to draw up a list of important decisions that need to be made — the ones I often avoid confronting. So I take time every Sunday to think about these tough issues and almost always make progress. Progress doesn’t always mean that you come to the final conclusion — sometimes it’s just a very rational next step that can put you on a path to the final decision.

#lessons learned from the founder of AppDynamics and Harness (Jyoti Bansal)

At my first startup, it took us 15 months and multiple pivots to find our PMF and ship a good v1 product. If we hadn't pivoted, we would have died. 

3 mistakes we made 👇 

1) Being too far ahead of the customer's current problems.

Our assumption: Everyone would be moving to the cloud in 2-3 years and would need a cloud monitoring solution. The reality was that it was a much longer timeline and there's no funding for being right eventually. 

Our pivot: Fixed our product so that it could be used even if a company wasn't on the cloud yet. 

2) Asking too much from our customers.

Our assumption: We thought our software would operate as SaaS, which meant that customers would have to adjust their approach to be compatible with our tech. But customers were telling us the opposite: they needed the entire solution to be on premise. 

Our pivot: Modified our platform so that companies didn't have to make any changes in application architecture to use our software. They brought their existing applications, and we would set them up. 

3) Trying too hard to sell our products instead of solving customer problems

Our assumption: We were designing a complex product that would manage the entire monitoring process — from finding to fixing issues. So that's what we kept selling. But customers only wanted the "find it" part. Our product would have been excessive. 

Our pivot: Cut back on our product roadmap and focused on simple application visibility. 

The hardest test for any startup is to find initial PMF. Finding it requires countless customer conversations and adjusting based on the feedback you get.

#ABL - Always Be Learning - today is your day to take the risk…Dr. Julie Gurner

Betting on yourself matters...because confidence is constructed. 

You create virtuous cycles or destructive ones, by the smallest of actions - your hesitancy, or your willingness to take risks. 

Here's how it works - from my chat with @ShaneAParrish

Construct Confidence.

More here...

Enterprise Tech

#cybersecurity consolidation continues (see my predictions from What’s 🔥 #380) - this time in cloud security - headline says $150-200M exit for Lacework which raised $2B but distributing $800M cash back to existing investors certainly helps. Last round was at $1.3B at $8.3B in 2021 so I’m sure in the heyday this was pari passu so everyone will get around 48 cents on dollar back. Still wild to think $1.2B was spent to get to $100M ARR. (CTech)

Cloud security decacorn Wiz is in advanced negotiations to acquire the American cloud security company Lacework, Calcalist has learned. The companies recently signed a Letter of Intent and are now in the midst of a comprehensive due diligence process, at the end of which a decision will be made as to whether the deal will go through. If completed, the transaction is valued at approximately $150-200 million, with Lacework’s $800 million in cash reserves to be distributed among the company's investors.

#more VC playbooks - this one, which has some pretty good data, from Menlo Ventures on Enterprise AI Apps - the real question is what do these numbers look like in the next 12-24 months when it comes to retention?

#Case in point - selling to prosumers and SMEs is brutal - Tome, a fast out of gates early winner in AI, cut 20% of its staff with a harder pivot to enterprise…question for many of these first to market AI native startups is how durable is the revenue as competition increases? What will churn look like? (The Information)

Another early leader among generative artificial intelligence startups has cut staff after struggling to make money. Tome, which helps customers use conversational AI to quickly create digital presentations, on Monday laid off 12 of its 59 employees, the company confirmed.

The San Francisco-based Tome is one of a number of generative AI startups, including copywriting firm Jasper and image generation startup Stability AI, whose momentum faded as buyers of AI software increasingly found cheaper alternatives or were disappointed with the technology’s performance. And because Tome’s product relied on AI made by other firms, it didn’t take long for competitors to emerge.

#yes, this is a thing, especially because of AI and need for data privacy and security which means on-prem is often better for regulated industries. Also check out Replicated.com/AI which allows SaaS vendors to distribute commercial AI applications to secure enterprise environments.

Multi cloud including on prem and collocation is the clear trend. Somewhat driven by the growth in AI inference and data gravity, 83% of enterprise CIOs in Barclays survey plan to repatriate at least some workloads in 2024, up from low point of 43% in 2020 H2.

Michael Dell

#🤯 the cost of cyberattacks is astounding - $872M charge in Q1 and up to $1.6B in potential damages to United Healthcare moving forward. Just to remind you, this was front and center in the earnings release - this is why cybersecurity revenue keeps going 📈 (Axios)

Between the lines: UnitedHealth said Tuesday that it expects $1 billion to $1.15 billion in "direct" costs in 2024 as a result of the attack. It projected another $350 million to $450 million "business disruption" hit, which includes lost revenue.

The incident caused a $872 million drag on earnings from operations in the first quarter. The exact full-year tally is uncertain in part because the company is still recovering from the attack, which led to hackers gaining access to a trove of sensitive information, including patients' hospital bills, financial documents and company contracts.

#In case you don’t know what an agentic workflow is, this 90 sec primer from Andrew Ng nails it - here’s the kicker, if you use GPT3.5 in an agentic workflow it performs better than just using GPT4 (zero-shot) - also nice shoutout to CrewAI (a boldstart portfolio co)

#more on AI and reasoning (FT)

Speaking at an event in London on Tuesday, Meta’s chief AI scientist Yann LeCun said current AI systems “produce one word after the other really without thinking and planning”.

Because they struggle to deal with complex questions or retain information for a long period, they still “make stupid mistakes”, he said.

Adding reasoning would mean that an AI model “searches over possible answers”, “plans the sequence of actions” and builds a “mental model of what the effect of [its] actions are going to be”, he said.

This is a “big missing piece that we are working on to get machines to get to the next level of intelligence”, he added.

LeCun said it was working on AI “agents” that could, for instance, plan and book each step of a journey, from someone’s office in Paris to another in New York, including getting to the airport.

#building Glean, ChatGPT for enterprise and some of the problems building the product (Practical Intelligence: Tamar, former CPO Slack, Venture Partner IVP and now Glean, President, Product and Tech)

Tamar: As you started building Glean, what were some of the problems you ran into?

Arvind: We underestimated how difficult it is to bring in data at each organization and integrate individual applications. Initially, we thought this would be easy given the fact that businesses have standardized the way they store information through SaaS APIs. But we soon ran into a number of issues. For one, it’s difficult to understand the governance behind data. Every SaaS application has built its own version of mission control permissions, which we replicate inside Glean’s search systems. This turned out to be a lot of work...

This is why we use a hybrid search system at Glean: We use traditional Information Retrieval (IR) techniques where we look at words in a query and match based on those words and their synonyms. But we also use AI to embed the queries and embed the documents to match semantically. Both of these factors come into consideration for Glean’s rankings, and the combination of the two yields better results.

For instance, if an employee is searching for a product roadmap within their company there are probably 10,000 documents that might come up. But you want the most relevant results for this employee. Not only should it be an authoritative source that’s used by the entire company, but it should be relevant to the employee’s team and include only the information they’re allowed to access within the company. This is where Glean differs from a traditional RAG, because we are a permissions-aware RAG engine using traditional keyword-based techniques, as well as modern embedding techniques within an enterprise.

#Gartner Says 75% of Enterprise Software Engineers Will Use AI Code Assistants by 2028

By 2028, 75% of enterprise software engineers will use AI code assistants, up from less than 10% in early 2023, according to Gartner, Inc. Sixty-three percent of organizations are currently piloting, deploying or have already deployed AI code assistants, according to a Gartner survey of 598 global respondents in the third quarter of 2023.

#The comprehensive AI Index Report for 2024 from Stanford Institute is out (summary from Luiza Jarovsky)

🚨BREAKING: The @Stanford Institute for Human-Centered AI publishes its Artificial Intelligence Index Report 2024, one of the most authoritative sources for data and insights on AI. Below are its top 10 takeaways:

1. AI beats humans on some tasks, but not on all;
2. Industry continues to dominate frontier AI research;
3. Frontier models get way more expensive;
4. The United States leads China, the EU, and the U.K. as the leading source of top AI models;
5. Robust and standardized evaluations for LLM responsibility are seriously lacking;
6. Generative AI investment skyrockets;
7. The data is in: AI makes workers more productive and leads to higher quality work;
8. Scientific progress accelerates even further, thanks to AI;
9. The number of AI regulations in the United States sharply increases;
10. People across the globe are more cognizant of AI’s potential impact—and more nervous.

Markets

#What ARR and growth does one need for an IPO? Alex Clayton from Meritech breaks it down with data here 🧵

Given that the IPO window could open soon, we thought it was the right time to analyze what it might take for a SaaS company to go public in today’s market. Moreover, there is a narrative that a business must be close to $1B of ARR to go public. We believe that is no longer relevant. For businesses that are $250M+ in ARR, growing quickly and durably, and with strong margins, now is the time to start thinking about the public markets. There is precedent for “smaller” IPOs that were massively successful in companies like CrowdStrike, Datadog*, HubSpot, Shopify, and others. Importantly, deciding to IPO isn’t just about financial profile, so we’ve included questions management teams should ask themselves to help determine if they are ready for the public markets.

#Big ideas from the 🐐 Vinod Khosla 👇🏼 - definitely worth a read

Entrepreneurs, with passion for a vision, invent the future they want. These are my predictions for abundant, awesome, technology-based, Possible Tomorrows (2035-2049) ... if we allow them to happen! #TED2024 
@TEDTalks

#🤔 - lots of investing in AI chip startups as of late but…Bojan Tunguz (ex-Nvidia)

OK, now that I’m out, I can finally say this publicly: LOL, no, sorry, you are not catching up to NVIDIA any time this decade.

#Goldman Sachs on AI - here’s CEO David Solomon in the earnings call - tailwind in financing for GS…along with developer productivity

I also want to touch on a topic coming up in virtually every client conversation I have, Artificial Intelligence. While there is broad consensus about the transforming potential of AI, there is an enormous appetite for perspectives on how certain aspects may play out, including the timeline for commercial impact, shape of potential regulations, impact on jobs, and where value will accrue in the ecosystem.

Today, we are proud to be at the forefront of advising clients on these topics and how to think about potential use cases in their operations. As we look longer-term, to the extent that this technology develops in line with expectations, there will be significant demand for AI-related infrastructure and as a result, financing, which will be a tailwind to our business.

For our own operations, we have a leading team of engineers dedicated to exploring and applying machine learning and artificial intelligence applications. We are focused on enhancing productivity, particularly for our developers, and increasing operating efficiency while maintaining a high bar for quality, security, and controls. Like with any emerging technology, a thoughtful approach and keen eye on risk management will be crucial.

#how to know if it’s you and not the macro environment (OnlyCFO)

Financial signs of weak product-market fit (PMF)

🚩When management has multiple signs over a longer period and they constantly blame macro conditions

More here...