Q2 earnings this past week were looming large as there has been lots of discussion lately on whether or not all of the infrastructure spend on AI will eventually translate into the next wave of spending on software. Here’s David Cahn doubling down on some comments this past week from Google and Meta CEOs.

This narrative has been playing out in the financial press and mainstream news as everyone has been waiting for any evidence of phase 2 of the AI buildout as we go from CapX on model training and infrastructure to software revenue. One of the bright lights in a brutal week for the market was ServiceNow, and while I agree we are certainly in a mini bubble now for AI, one which will ebb and flow, it’s great to see companies like ServiceNow executing. While still early, and I’d argue late 1st inning, here’s more from Bill McDermott on how ServiceNow crushed the quarter (stock was up 15% day after the call).

Our Q2 results prove that we are delivering elite-level execution. Subscription revenue grew 23% year-over-year at constant currency, which is 100 basis points above the high end of our guidance. CRPO grew 22.5% at constant currency, which is 200 basis points above our guidance. Operating margin was over 27%, nearly 250 basis points above our guidance.

What's the headline? ServiceNow is relevance is the AI platform for business transformation is soaring. This is a growth company on an unprecedented trajectory. Moving to the topic of AI. Now Assist net new ACV doubled quarter-over-quarter, significantly overachieving our expectations, and it became the fastest-growing new product in the company's history.

We signed 11 Now Assist deals with $1 million-plus NNACV in Q2, two of which were over $5 million. We had impressive wins across industries, from banking, healthcare and manufacturing, to semis, tech and many others. As an AI Lighthouse customer, Stellantis is using ServiceNow as its strategic AI platform to manage operations across HR, finance and supply chain. American Honda selected Now Assist as their GenAI solution of choice to improve deflection and efficiency of service, delivering a world-class experience to their employees.

So that’s $11M from Now Assist and another $10M from 2 x $5M deals and you have $21M this quarter from AI, and we can expect even more growth as Bill calls it “the fastest-growing new product in the company's history.” Furthermore, Bill went on CNBC (watch around 4’20”) to say that the pipeline is up 50% YoY because of GenAI with an added $1B of revenue opportunities because of GenAI.

Here’s Bill McDermott sharing the big vision in the most recent earnings call.

"Our relevance as the AI platform for business transformation remains stronger than ever as CEOs are looking for new vectors of growth, simplification, and digitization," said ServiceNow CEO Bill McDermott. "ServiceNow intends to reinvent every workflow, in every company, in every industry with GenAI at the core."

Which sounds much like Satya’s quote from October of 2022 where he said AI would be part of every product, and it actually is!

Despite all of the flash around AI, let’s not forget 👇🏼 what you see below - that is the main UI for the product! It goes to show that ultimately what really matters is having a killer enterprise sales execution machine which will get you really big over a long period of time! 🤣

Finally, I wanted to note this when it comes to LLMs, hallucinations and cost.

Do we have the big idea? ServiceNow is putting AI to work for people. Our GenAI strategy is focused on infusing intelligence into the flow of work, end-to-end across the enterprise, every department, every persona. With our native integrations, we already help people orchestrate across different systems and data sources. Now we can train the machines to do the low-value work so the people can up-level to the knowledge work.

The second pillar is viability, does our approach to GenAI make business sense? We set the tone on innovating with domain-specific, multimodal language models. This approach is more accurate. It doesn't have the hallucinations of the large models from the Internet. It's fast to the train, and it's faster to deploy these models.

It's more cost-effective and sustainable because these models are less GPU hungry than public models. The final pillar is feasibility. Can we actually do it? This is the best pillar of all for ServiceNow because we've actually done it. When you look at our own deployment of Now Assist, the initial results are staggering.

Repeat after me, domain-specific multimodal language models vs. general purpose large models are what works for enterprise sales.

Switching gears, here are 2 more examples of GenAI in the real world as banks like JPMorgan Chase and Morgan Stanley began rolling out custom solutions.

First, JPMorgan pitches in-house chatbot as AI-based research analyst as about 50,000 staff are using a large language model designed by the bank to boost productivity (FT).

Second, as per the WSJ:

“Morgan Stanley said it finished deploying its second generative artificial intelligence application to financial advisers last week, favoring homegrown solutions over out-of-the-box tooling from tech providers.

The new tool, AI Morgan Stanley Debrief, which summarizes video meetings and generates drafts of follow-up emails based on them, was Morgan Stanley’s second generative AI use case built in collaboration with OpenAI. It follows the September 2023 rollout of its AI knowledge assistant tool that helps financial advisers quickly track down information from Morgan Stanley research.”

Finally, I’m super excited about the launch of Noded AI which my firm boldstart ventures led which included others like Bessemer Venture Partners, 20VC Growth, Stewart Butterfield (co-founder Slack), Tamar Yehoshua (CPO Glean), Joe Teplow (SVP Salesforce Labs) and some others.

Noded AI is an AI-native notes workspace which makes the notes we all take the center of the enterprise - just work, take notes, and we bring the system to you giving you superpowers on customers, prospects, etc powered by integrations in lots of other SaaS apps. Visit Noded and check out the video.

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

Scaling Startups

#dream the dream, also took 2 year to get over $1M Revenue and 3 to go over $5M - also shows the value of compounding growth even at 50% as the company hit $78M Revenue

#equally applies to the founder journey “Albert Camus on Writing and the Importance of Stubbornness in Creative Work”

To someone who asked Newton how he had managed to construct his theory, he could reply: “By thinking about it all the time.” There is no greatness without a little stubbornness.

Nearly a century after Tchaikovsky asserted that “a self-respecting artist must not fold his hands on the pretext that he is not in the mood,” Camus adds:

"Great novels… prove the effectiveness of human creation. They convince one that the work of art is a human thing, never human enough, and that its creator can do without dictates from above. Works of art are not born in flashes of inspiration but in a daily fidelity."

#All the startup accelerators you can apply for (Andrew Yeung)

1. PearX ($250k-$2m)
2. Accel Atoms ($500k)
3. Antler ($200k for 8%)
4. Soma Capital ($100k)
5. Sequoia Arc (Variable)
6. a16z Speedrun ($750k)
7. LAUNCH ($125k for 7%)
8. OpenAI Converge ($1m)
....

#lessons learned from Notion cofounder Ivan Zhao as it scaled to 100M users and they wandered around the desert for 5 years till launching on Product Hunt in 2018! Similar to the Clay story…

Last week, @NotionHQ passed 100M users!

I found the photo I took after we reached 1000 users.

And I still recall the day when we passed 1M in 2020 right before covid.

I feel grateful, introspective. Looking back, there's so much learning, yet much remains unchanged...

1/n

In our early years, we were rather lost. 

We started in 2013, but it wasn't until 2018 w/ Notion 2.0 
@ProductHunt launch that we saw signs of traction.

We had no business sense, struggled with building a horizontal tool. Notion almost died (thanks for the bridge, mom!)

More here...

Enterprise Tech

#AI models are running out of data to train on but synthetic data is not the silver bullet, new paper from Nature going viral - here’s what the FT wrote about it

The use of computer-generated data to train artificial intelligence models risks causing them to produce nonsensical results, according to new research that highlights looming challenges to the emerging technology. 

Leading AI companies, including OpenAI and Microsoft, have tested the use of “synthetic” data — information created by AI systems to then also train large language models (LLMs) — as they reach the limits of human-made material that can improve the cutting-edge technology.

Research published in Nature on Wednesday suggests the use of such data could lead to the rapid degradation of AI models. One trial using synthetic input text about medieval architecture descended into a discussion of jackrabbits after fewer than 10 generations of output. 

The work underlines why AI developers have hurried to buy troves of human-generated data for training — and raises questions of what will happen once those finite sources are exhausted.

Alexandr Wang from ScaleAI chimes in on X

3/ This core idea is very important to pay attention to:

Synthetic data can create a short-term boost in eval results, but you will pay for it later with model collapse!

You accumulate debt with mangling the model that starts invisible, and is very hard to repay...

7/ watch this space closely.

My prediction is that model developers who are not careful about training on synthetic data with no information gain will find their models getting steadily stranger and dumber over time.

Synthetic data accumulates a debt with the model that must be repaid at some point.

#🤯 OpenAI burning an estimated $5B a year as open source is rapidly commoditizing LLMs

#what is value of closed source as open source LLMs are catching up so quickly and rapidly commoditizing the billions of 💰 spent to train OpenAI and other models

#Cybersecurity 💪🏼 growth round for Vanta (SOC 2 Compliance), up from $1.6B post money June 2022

Thrilled to announce we’ve raised $150mm Series C at a $2.45bn post valuation led by Sequoia Capital alongside all our existing investors – Y Combinator, CrowdStrike, Atlassian, HubSpot, Workday, and Craft Ventures – and new investors Goldman Sachs and J.P. Morgan. The terms on this round are exactly as we aim Vanta experiences to be: straightforward and clean.

I couldn’t be more proud of how we’ve put customers first since the Series B two years ago. That daily work shows up in this round and in our output metrics: Vanta now has over 8000 customers, almost 2x as many as last year – we’re adding customers faster than ever before; our customer growth outside of the US is faster than inside; and Trust Management is starting to catch on.

We didn’t need to raise this round – we had ⅔ of the Series B in the bank – but we chose to in order to better serve our customers:

🤖 We’re building AI and LLM-enabled features for our customers first, not for ourselves. Questionnaire Automation is actually really good – 80-90% answer approvals! – and Trust Center chat quite useful. Smart Recruiters saves 20 hours each week with a streamlined security review process.

#cybersecurity on 🔥 with another growth round as Chainguard in the supply chain security space raises $140M Series C round at a $1.14B valuation led by IVP, Lightspeed, Redpoint

Chainguard triples valuation to $1.12 billion in less than one year, as its secure-by-design, minimalist approach to container images becomes a de facto technology for locking down open source software in the modern software supply chain.

#Morgan Stanley CIO Survey from 2 weeks ago - one 🔑 note on AI as we enter Q2 earnings season where many on Wall Street want to see impact of AI beyond CAPX spending

2Q24 data indicates a push-out in CIOs' expectations for Generative AI projects to enter production, with 26% of CIOs expecting first projects to enter production after 2025, an uptick from 19% of CIOs over our past two surveys (Exhibit 32). 

The longer ramp period may suggest a lack of conviction around GenAI decision making for CIOs, as a solid reference architectures aligned with security and data protocols are taking longer to materialize than initially expected. 

However, our survey contained bullish GenAI data points as well, with net AI/ML prioritization continuing to increase (+267 bps net prioritization from 1Q24), underscoring the burgeoning interest and excitement from CIOs surrounding these innovative technologies, Exhibit 4. 

Net, 2Q24's indication of stable budget growth expectations and positive revisions across sectors supports the durability of IT spend. However, with moderation in the up-to-down ratio and the push-out in Generative AI project implementation time line potentially indicative of GenAI induced 'analysis paralysis', onus shifts to 2H24, as enterprises share spending intentions and the ability for GenAI to inflect budgets more materially in 2025.

#true to an extent…read comments but point is Claude Sonnet 3.5 is pretty damn good!

#legal tech AI growth round - Harvey, AI for Lawyers, raised $100M at a $1.5B valuation with crazy growth (Sheel Mohot)

They say they tripled ARR from December, implying that they are at $30M ARR today <2 years from founding.

Anonymous internet commenters say it doesn’t work but inside investors keep putting more money in 🤔

read 🧵 for more...

#great report from Coatue on the future of robotics - super promising but it won’t have its “ChatGPT moment”

#Crowdstrike fallout continued - what will Microsoft do, lock 3rd party vendors out of OS and then be only solution to go deep into kernel to secure endpoints? (Cloudflare co-founder and CEO, Matthew Prince)

Here’s the scary thing that’s likely to happen based on the facts of the day if we don’t pay attention. Microsoft, who competes with 
@CrowdStrike, will argue that they should lock all third-party security vendors out of their OS. “It’s the only way we can be safe,” they’ll testify before Congress.

But lest we forget, Microsoft themselves had their own eternal screw up where they potentially let a foreign actor read every customer’s email because they failed to adequately secure their session signing keys. We still have no idea how bad the implications of #EternalBlue are.

So pick your poison. Today CrowdStrike messed up and some systems got locked out. That sucks a measurable amount. On the other hand, if Microsoft runs the app and security then they mess up and you’ll probably still be able to check your email — because their incentive is to fail open — but you’ll never know who else could too. Not to mention your docs, apps, files, and everything else.

Today sucked, but better security isn’t consolidated security. It isn’t your application provider picking who your security vendor must be. It’s open competition across many providers. Because CrowdStrike had a bad day, but the solution isn’t to standardize on Microsoft.

And, if we do, then when they have a bad day it’ll make today look like a walk in the park.

#btw if you think the easy answer is for Microsoft to wall off its OS here’s also why it can’t

CrowdStrike’s bug was so devastating because its security software, called Falcon, runs at the most central level of Windows, the kernel, so when an update to Falcon caused it to crash, it also took out the brains of the operating system. That is when the blue screen of death appeared.

In 2020, Apple told developers that its MacOS operating system would no longer grant them kernel-level access.

That change was a pain for Apple’s partners, but it also meant that a blue screen-style problem couldn’t happen on Macs, said Patrick Wardle, the chief executive of Mac security maker DoubleYou.  

“What it meant was that a lot of third-party developers, ourselves included, had to rewrite our security software,” he said.

A Microsoft spokesman said it cannot legally wall off its operating system in the same way Apple does because of an understanding it reached with the European Commission following a complaint. In 2009, Microsoft agreed it would give makers of security software the same level of access to Windows that Microsoft gets.

#🤯 CrowdStrike global outage to cost US Fortune 500 companies $5.4bn - Banking and healthcare firms, major airlines expected to suffer most losses, according to insurer Parametrix (The Guardian)

#How a North Korean Fake IT Worker Tried to Infiltrate Us (KnowBe4) - fascinating and I highly encourage you to read it

TLDR: KnowBe4 needed a software engineer for our internal IT AI team. We posted the job, received resumes, conducted interviews, performed background checks, verified references, and hired the person. We sent them their Mac workstation, and the moment it was received, it immediately started to load malware.

Our HR team conducted four video conference based interviews on separate occasions, confirming the individual matched the photo provided on their application. Additionally, a background check and all other standard pre-hiring checks were performed and came back clear due to the stolen identity being used. This was a real person using a valid but stolen US-based identity. The picture was AI "enhanced". 

The EDR software detected it and alerted our InfoSec Security Operations  Center.

#‘I Need to Identify You': How One Question Saved Ferrari from a Deepfake Scam (Bloomberg)

* Benedetto Vigna was impersonated on a call using AI software
* Large companies are being increasingly targeted with deepfakes

#🤔 pictures with deep meaning (Moral Wisdom)

Markets

# wit da Wiz or witout da Wiz (this is for those Philly Cheesesteak fans) - no deal with Google for $23B, Wiz walks away citing market validation after the leak/offer - here’s the letter to employees from Assaf, co-founder and CEO (TechCrunch).

#🤔 Why is Xi Jinping building secret commodity stockpiles? (The Economist)