Clara Shih

We're launching the inaugural edition of Beyond Benchmarks, a deep dive into the metrics and trends observed across the early-stage enterprise cloud market. This is the first look at some of the business data around GenAI in SaaS. Some initial observations: 1) 60% of SaaS companies have already integrated GenAI into their offerings (60% of these as part of their current product and 40% as a new product), and an additional 20% are planning to do so this year. Almost half aren't monetizing these features. Those that are are primarily experimenting with both usage-based and flat fee pricing models. 2) Most companies are utilizing OpenAI as their primary LLM, but many are experimenting with multiple models - and we are noticing an early trend towards intelligently routing GenAI inference requests to different models based on cost, performance, and security. 3) Though we are in the initial innings of GenAI, we are seeing some positive signals - companies that implemented GenAI had 7% higher NDR than those that did not. Still early days, but we'll be doing this benchmarking report regularly to track and share how things progress in this new era. Link to report in comments.

104

6 Comments

This is a worthwhile overview on the bets Satya Nadella has made at Microsoft. His overarching them has been that with AI, every old fight and every old bet the Microsoft has made was back on the table, is paying off. He's made some big bets, including in OpenAI. A few didn't pay off, and a few more are still up in the air. He's successfully put Alphabet Inc./Google back on it's heals. He's forced Cisco to step up their security game. It's fascinating to see how he's revitalized the company and made it relevent again while building its resiliency to weather a few failures along the way. #Microsoft #satyanadella #ai #innovation #competition #technology https://lnkd.in/e7ssR6JM

2

AI memory is fundamental for conversational AI and AI agents. It is critical for building personalized experiences and improving reasoning. Check out Letta’s course to understand how to use AI memory to enhance the customization and quality of your AI app.

15

Service-as-software is a $4.6T opportunity. But no one knows how to actually grab that opportunity yet. Enterprise software has gone from assisting human workflows to now, LLMs giving software the ability to actually perform parts of the workflow. Instead of relying on timely and sometimes inaccurate human inputs of structured data, today's systems actively capture and process both structured and unstructured data. The service-as-software shift was huge—but we're on the edge of one that's even bigger. My Foundation Partners Joanne Chen and Jaya Gupta have now outlined how to put the service-as-software paradigm shift into action: through a System of Agents, which doesn’t just streamline a workflow but actually completes whole tasks, becoming the worker. A System of Agents mirrors a human team, with agents taking on collaborative, specialized tasks and continuously learning from each other, just as human teams do. Advice for founders building with Systems of Agents: → Position at the data source, own the path. Traditional software waits for humans to input data, whereas Systems of Agents capture it at the source. When an agent owns the interface where data is born, the system can then orchestrate every downstream action. → Move beyond software budgets, tap workforce spend. When a System of Agents completes a whole job task, it should be categorized as a personnel cost, not a software expense. → Use AI’s 24/7 nature and scalability to create new workstreams. AI agents working in areas like healthcare can operate 24/7, providing reliable support and an around-the-clock presence. → Stay ahead of pricing and business model shifts. Service-as-software shifts the core business model from seat-based pricing to outcome-based pricing; value is placed on the tangible improvements that AI delivers. In the past 6 months, we’ve heard from hundreds of startups working in this space. After those conversations, we’re announcing the System of Agents 50, which includes the companies you see here. 👇 Full blog in the comments. 

736

65 Comments

Perhaps lost in the news from OpenAI Google this past week was this new model from Microsoft Phi-4. Only 14B parameters but and on par with much larger Llama and Qwen models on MMLU. Out performing Gemini 1.5 Pro on math competition problems. https://lnkd.in/eUPFXERp

30

AI, Longevity and Love - how are they all connected? They make each other better and are essential to each other’s success Lots of fantastic discussions about how AI will bring out the best in humans including augmenting our time and ability to connect at DOC in Napa and Wisdom 2.0 in SF this past week. Huge thanks to Jordan Shlain, MD, Kevin Ryan and John Battelle for organizing leaders across traditional medicine, longevity, academic research, and biotech and Soren Gordhamer and Mark Hyman, MD for bringing together the best of AI and longevity AI shows incredible promise for healing humans and increasing longevity, here are my key takeaways from these events: We can live to 95 today if we focus on longevity and healthspan, there are biotech solutions using AI to epigenetically reprogram cells to reverse aging and disease that will enable us to extend lifespan by 20+ years but it will take for these to be real Fun to see Mark Hyman, MD interview Eric Verdin, they both discussed how they lowered their biological age by 10+ years and feel the best ever There are so many simple things we can do, walking 30 minutes a day reduces cardiovascular, alzheimer’s and cancer risk by 40% but 80% of the US doesn’t even do this Within 5 years we will have a whole new set of FDA approved psychedelic drugs for depression, anxiety, PTSD, etc. Immunotherapies will also expand so we can avoid chemo and living cancer free will be in reach We need the US to move to a functional medicine, integrative care model and get out of this sick care model, this level of innovation is more likely to come from upstarts, the incumbents have too much revenues to lose in this transition AI will not replace physicians, instead those physicians that don’t use AI will be replaced. 50% of physicians are burned out, AI can help! Fantastic connecting with so many brilliant minds, including Mark Hyman, MD, Eric Verdin Neal Khosla Laura Esserman Abby Levy Sheila Lirio Marcelo Leonard Zon Arjun Desai David Eagleman Lois Quam Vinod Khosla Emi Gal and many others #healthcare #longevity #ai

60

4 Comments

Thanks for sharing these insights. What I am seeing is that orgs test an idea using a proprietary #LLM like #OpenAI, then put it into production with an open source LLM using a development platform like #Llamaindex. That way, private data remains private.

1

To grow an AI application, making it part of users’ workflows might be just as important as the tool you’re building. Microsoft's Chief Strategy Officer Bobby Yerramilli-Rao crystallized this insight in our recent conversation: "The technology and speed of writing applications will far outpace people's ability to change their behavior." Video conferencing offers a perfect case study. The technology existed for years with modest traction. Then COVID hit, forcing remote work into standard operating procedures. Now that video calls are embedded in daily routines, adding AI features becomes natural because users don't need to learn new workflows. This principle extends across enterprise software. Technical founders often believe superior technology will drive adoption. But in my experience professionals rarely abandon established processes, no matter how compelling the alternative. AI adoption starts with augmenting existing behaviors, not disrupting them.

93

8 Comments

Generative AI: miracle worker or sophisticated guesser? While AI can now write, code, and create, it's essentially still a prediction machine. This revelation challenges how we view and utilize AI in business. What does this mean for your AI strategy? Let's discuss. #ArtificialIntelligence #BusinessStrategy #TechInnovation #AILimitations #FutureOfWork

2

In a recent Office Hours, Baris Gultekin, Head of AI at Snowflake & a friend from Google, shared his insights into Snowflake’s approach to AI. It was a great conversation to understand how one of the leading data companies is pushing AI forward. The video for the session is here : https://lnkd.in/gfBesq35 & the audio is here : https://lnkd.in/gUK8hR66. There are four major pushes for Snowflake within AI. 1. Bringing AI to Data with Cortex. Cortex is a suite of AI building blocks that enable customers to leverage large language models (LLMs) & build applications. Both structured data & unstructured data (extracting information from presentations & pdfs) can be built using Cortex. 2. Developing open source initiatives including a data catalog, Polaris, & an open LLM Model Arctic which focuses on SQL performance. Open source is a new motion for Snowflake. 3. Securing data sharing to bring data into customers’ accounts for AI. Data sharing across customers enables a marketing organization to use data from a credit card company to build a model for customer marketing. 4. Developing a platform atop Snowflake with native apps to enable a retailer to provide a search experience to their wholesalers, for example. We also chatted about the importance of data access security for AI applications. For many data leaders, this is one of the top priorities to understand before deploying AI. Thanks, Baris, for sharing your perspectives!

31

4 Comments

#AI has become such a pioneering field that it now headlines The Nobel Prize results, marking a historic moment for AI in both physics and chemistry. Geoffrey Hinton and John Hopfield have been awarded the 2024 Nobel Prize in #Physics for their foundational work in artificial neural networks, which are now driving innovation in machine learning. Their breakthroughs have revolutionised how we interact with technology, enabling advancements from language translation to image recognition and even aiding complex physics research. Simultaneously, Sir Demis Hassabis and John Jumper of Google DeepMind have won the Nobel Prize in #Chemistry for their work on AlphaFold, a groundbreaking AI system that predicts protein folding with unprecedented accuracy. This achievement is transforming the fields of biology and medicine, accelerating drug discovery and deepening our understanding of molecular structures. These awards highlight the immense impact of AI across scientific disciplines, proving that machine learning is not just a technological advancement—it’s at the forefront of modern scientific discovery. #NobelPrize2024 #AI #MachineLearning #Physics #Chemistry #DeepMind #AlphaFold #ScientificBreakthroughs #FutureOfAI

10

I’ve been fortunate enough to do AI shows with Sam Altman, Yann LeCun & Arthur Mensch. But this show today with David Cahn at Sequoia Capital is the best of all of them: 🧠 AI’s $600BN question 🏃 Why no one will never train a frontier model on the same data center twice 🏗️ Servers, steel & power: The 3 pillars of the future of AI Not one to be missed, my 8 key takeaways 👇 1. The Oligopoly That is Blowing Billions on AI - Microsoft, Azure, & Google now represent $7T market cap. - Of course they will spend billions to protect their oligopoly - What is most interesting is Meta is the only incumbent that is financing this not with cloud being their cash cow. 2. How Startups Could Win with the Cost of Compute Coming Down - Big tech companies are producers of compute & startups are consumers of compute. - Overproduction of compute = lower prices.  - This gives startups higher margins & makes them more valuable. 3. Is Compute the Currency of the Future or Do the Costs Come Down and It Is Commoditised? - We are forgetting “compute” is a euphemism for a $2BN data center in the middle of nowhere. - They have GPUs & liquid cooling systems that need to be constantly optimized & upgraded. - We cannot simply build 15 years worth of compute all at once.  No one's ever going to train a frontier model on the same data center twice. 4. Data Centre Teams and Model Teams Need to Be Coupled - You cannot have separate teams running data centers & building models. - Elon & Zuck control all their data centres. - It will not work as these models get better & better. 5. Why the Data Center Is the Most Important Asset - There is not much difference between these models today. - AI researchers are jumping from one lab to another. - Scaling laws will become more dominant as these models get bigger. 6. Servers Steel and Power: The Three Bottlenecks of AI Three things: 1. Servers: NVIDIA, AMD, Broadcom… 2. Steel: Construction & real estate developers for data centers. 3. Power: All the energy required for AI. - More factories will be built in 12 months for AI and we will see another industrial revolution. 7. You Have to Have a Cash Machine to Play the Big Model Game - Amazon has AWS - Meta has Instagram - Microsoft has Azure - You need a cash machine to compete and it CANNOT be the AI business. 8. There is One Definition of Success in Venture: Are you a Slugger? - If you are generating billion dollar gains, you are a slugger. - Doug Leone, Pat Grady, Andrew Reed are all sluggers. - You cannot be good at this business without being a slugger. (links in comments) #founder #funding #business #investing #vc #venturecapital #entrepreneur #startup #seed #funding #sequoia

190

26 Comments

"Software is eating the world. And AI is eating software." —Jensen Huang, CEO of NVIDIA (2017) Over the past 4 months, I’ve helped implement AI solutions at 8 different companies to improve their GTM strategies. Based on that experience, I’m convinced that AI will completely change GTM in 2025. Here are the four most common ways I use AI: Example 1 → LEAD SCORING Think: a more personalized MadKudu (sorry MadKudu). Tailored *exactly* to your needs. This is stuff you CAN'T get in a standard B2B database, eg - SaaS? Funding >$5M? Public or Private? Hiring x department? Mentions of x technology in job posting? Has a generative ai initiative right now? Following the company's founder on LinkedIn? Open-sourced technologies in use. Etc. (The only limitation here, is your imagination) The legend, Andreas Wernicke, helps me build all of this in a centralized Clay table — so, it can then be rolled up into custom-built logic to tier out companies as Tier 1, Tier 2, Tier 3, or Unqualified. (If you prefer, you could do a—weighted—score instead) It's magic. Relatively cheap. And 100% customizable to your business. Example 2 → MICRO-CAMPAIGNS I wrote a dedicated post on this subject: https://lnkd.in/erSS5Kbm I'm seeing this motion quietly, but quickly, become one of the most important ways of generating pipeline in 2024. Here's how I define "micro-campaigns": - Small, highly-targeted lists of, say, 50-250 contacts - These lists are generated using *very* specific filtering and/or signals (usually using AI/agents at least for certain parts) - Each list (aka "micro-campaign") is relevant this week, but won't be next month I can't come up with these ideas for you (you are the expert). All I can do is pull these ideas out of you and scale them (using AI). Example 3 → PLG ENRICHMENT+AUTOMATION Any sign-up gets automatically enriched in the background with "rich data" that is highly relevant to the company. AI agents don't sleep. Use them. Example 4 → CRM ENRICHMENT Most B2B databases have "standard" data in them, like company size, location, headcount by department, funding, etc. What you really want in your CRM is the "non-standard" stuff (aka "rich data."). This is where the real alpha is. Ex: Tech stacks that are hard to find (use Sumble). Initiatives that don't have a keyword in a database for a specific upcoming project, but an LLM can scrape the latest earnings calls. Mentions on social of keywords. Closed-lost deal notes summarized. Etc. You need these data points in your CRM for every target account, so your GTM team is always up to date. –––––––––––––––––––––– PS - I write a weekly newsletter on the intersection of AI and GTM. Get the next post in your inbox by signing up here (it's free): www.thesignal.club

27

12 Comments

At a recent Adobe earnings call, Goldman Sachs analyst Kash Rangan asked: „Will generative AI be so good that it’s the end of the creative process—so we don’t need creative folks?“ Adobe‘s CEO Shantanu Narayan replied that the company expects an expanded customer base; the message to investors: with AI, everyone becomes creative, not „only“ designated designers, developers and artists, scientists and storytellers, performers and composers. (Nota bene, Kash reiterated a Buy rating on Adobe afterwards.) But of course, the question posed is not only key for investors, or for creation enablement companies such as Adobe, or for the wider public, but also, naturally: for creative technologies industries. More creators, that‘s fantastic empowerment—but: will creative professionals become extinct? Or less dramatic: what will creative technologists be doing in the foreseeable future? Let’s look into the history of computing to perhaps shape up new perspectives: - borrowing heavily from Xerox / Xerox PARC computers, Apple with the #Lisa had brought the Graphical User Interface (incl. direct manipulation, drag and drop, double clicking to launch, moving and resizing of windows) to the PC mass market in 1983, followed by the original, way more successful Macintosh -> this ignited the Desktop Publishing market and Human-Computer Interaction and later, with #GUI based software, boosted all professional content & product industries, including music, film and game production, as well as work in Human-Machine Interfaces and well beyond. - The DIY / Do-It-Yourself idea, spanning #hacking & the #maker culture, the game #demoscene, #IndieGames, #PunkRock and #ElectronicMusic, to #podcasting and #videos on social media, and #StartUp culture, too, has been powered and profited from computational creativity, oftentimes not driven by commercial interests, but by the wish to express and nail an idea. Given this: I think that with AI becoming the keystone of the creative process, we will see a culture of ‚more‘— both: more crap; more fake; more makers and outcomes and processes that are (from a touch to -almost- 100%) artificial. and: more innovation; more beautiful, „living“ outcomes; more unexpected edges; more DIY; more curating; more orchestrating with AI being a key component of the overall composition. But will we have more time, too? Isn’t less more? How do we pertain our humanity? How do we employ AI and creativity to build a better world? Someone recently said that they don‘t want an AI to create songs and videos, but that they want an AI to clean the kitchen so that they can be creative. What do you think? #AI #GenAI #Creativity #CreativeTechnologies #ComputationalCreativity #Innovation

16

4 Comments

AI agents are evolving rapidly, merging reasoning with tool-calling abilities to tackle complex tasks. This fusion is reshaping how AI interacts with its environment, from web searches to API calls. But it's not without challenges – balancing multiple tasks and selecting the right tools are critical hurdles. As the field progresses, expect to see more specialized models emerge, each fine-tuned for specific functions like planning or tool execution. #AIAgents #GenerativeAI #MachineLearning #TechInnovation #FutureOfAI What specialized AI models are you most excited to see develop?

5

PDFs + LLMs will be the QR code of B2B. So much of business-to-business interactions involve data trapped in PDFs — medical data, government data, transaction data. Even though it is "digital", it isn't easily computer readable, until now. Large language models make it easy to convert this into nicely structured data that can be used for decision-making. This means that hospitals, governments, real-estate developers, industrial farms, etc can continue publishing information in PDFs, and the consumers of that information will use LLMs (or apps built on LLMs) instead of painstakingly reading the PDF.

204

10 Comments