Hot in Enterprise IT/VC - What’s 🔥 in Enterprise IT/VC #384
Another week has gone by, and yet another LLM was released surpassing existing industry benchmarks - this time it was Claude from Anthropic. These are pretty 🤯 numbers. Opus is for Claude Pro users, Sonnet remains free, and Haiku is the “fastest and most cost-effective model on the market for its intelligence category. For the vast majority of workloads, Sonnet is 2x faster than Claude 2 and Claude 2.1, while Opus is about the same speed as past models.” As I’ve written in the past, the infrastructure layer is moving so fast that first mover is not necessarily an advantage. In addition, many founders that are building now are thinking about how they can swap today’s OpenAI model, for example, for what may be better in the future - just wait, OpenAI’s next model is supposed to be even better. With Anthropic’s most recent release, there has been lots of discussion around where value accrues in the AI stack - LLMs, infra, apps. Marc Benioff from Salesforce chimes in here
I also suggest checking out this 🧵from Jason Warner, ex-CTO of Github, and co-founder/CEO of PoolsideAI, building a new foundational model for code along with apps on top, it’s CoPilot…
I really like Jason’s thinking around full stack foundation models which includes the LLM and the apps on top:
If one can build a domain specific LLM with access to data that others cannot get with the idea of building focused apps on top of your own specific model like code or robotics, then one will have the ability to create significant value. This model is not for the faint of heart as training LLMs costs significant 💰 and Poolside, for example, has already raised a $126M seed round 😲. Next up is the infrastructure layer which includes all of the tools that infra folks and developers will use to build these applications and deploy them to the real world. This includes your traditional cloud vendors like Microsoft and AWS Bedrock who are best positioned to win here, 37 AI observability tools we’ve already seen, frameworks to build and deploy agents, 27+ vector databases...you get the point. There will be some huge winners here for sure as incumbents like MongoDB may find it hard to get the developer traction needed for this new category, but man, I’ve seen mountains of 💰poured into this space that it’s hard to tell who will win. In addition, when you have the foundational model companies like OpenAI offering its own version of fine tuning in a simple framework, one has to wonder what else every other LLM provider will offer to make it so easy for devs to call an API and build and deploy. Buyer beware in this category. Last but not least we have the product folks, yes the products themselves, and I’m not talking about thin wrappers and beautiful demos, but the products which just blow people away. Here are my thoughts which, btw, I’m sure will change again in the next few months as this market is moving so fast! First thing we need to remember is that AI is a technology and not a product. Zuck explains it well from Meta’s most recent earnings call transcript:
So where in the stack is less important for me vs. thinking about products and the user and how the initial experience leads to a LT advantage. Whether LLM, infra or app, show me a product that feels magical 🪄, that feels 10x better to the end user whether a dev tool, infra monitoring, cybersecurity or app, and you have a chance. Incremental won’t do - and yeah that magical experience is likely powered by a data moat but also involves the basics - UI, UX, data, etc. Some examples include Perplexity in search (not enterprise but illustrates my point) - it is blazing fast and works, Superhuman - simplest is often the most powerful use cases of AI (see video from What’s 🔥 #383 last week and think about the amount of email data it is sitting on to create personalized responses (also proud to be first investor), LangChain and CrewAI also have amazing developer experiences. With respect to Crew, it’s a super easy and powerful AI agent framework (video of how easy to build a crew of financial analysts) that’s attracting a ton of attention. What's the cybersecurity equivalent? The Datadog equivalent? There are so many amazing, talented founders starting new companies, and IMO there is no better time to start a company and "Inception Invest" but let’s change the conversation from a macro one of where value accrues in the AI stack to the micro and first principles thinking! What problem are you solving, for who, and how are you doing it uniquely to make an end user’s life 10x better with your product than without? Answer the “why now” - what can exist today that previously could not because of these new LLMs? What is your data acquisition strategy + advantage? How will you outmaneuver incumbents and how will your product be built to last? On the latter, there has also been a ton of debate. Incumbents have the data and distribution advantage, and Microsoft is crushing it but look at Google. Did you see that Gemini launch - what a disaster! Not every incumbent is like Microsoft - some like Google can’t get out of their own way when launching one of most important products in recent history. There are no absolutes! If the Internet felt like 🐶 years, then AI feels like 😽 years (first 2 human years = 24 cat years). It’s time to compete, to have some fun, and to deliver some magical experiences that aren’t just demos but go much deeper. It’s about solving the problem for the end user and making that experience magical. There will be lots of losers but some insanely big winners as well. Don’t think stack first, think user first. The good news is that the universal principles of success always matter - founders, team and execution always win, and this is no different in 2024 wherever you may roam on the AI stack. Like the Metallica song remember this: As always, 🙏🏼 for reading and please share with your friends and colleagues. Scaling Startups
Enterprise Tech
Markets
What's Hot in Enterprise IT/VC is free today. But if you enjoyed this post, you can tell What's Hot in Enterprise IT/VC that their writing is valuable by pledging a future subscription. You won't be charged unless they enable payments. |
Older messages
What’s 🔥 in Enterprise IT/VC #383
Saturday, March 2, 2024
The short version of what LPs and VCs are thinking - notes from an AGM... ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
What’s 🔥 in Enterprise IT/VC #382
Saturday, February 24, 2024
Platform vs. best of breed in cybersecurity (PANW) + the rise of AI Security + AISPM, yes, we do need another security posture management category 🤣 ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
What’s 🔥 in Enterprise IT/VC #381
Monday, February 19, 2024
Net Dollar Retention Update - what's best in class post-ZIRP? ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
What's 🔥 in Enterprise IT/VC #380
Saturday, February 10, 2024
What to expect for cybersecurity startups in 2024 when 4 companies represent $300 Billion of market cap + 5 account for >51% of all revenue...
What's 🔥 in Enterprise IT/VC #379
Saturday, February 3, 2024
Microsoft delivering on AI + thinking about second order effects for startups in a world where 50% or more code is written by AI?
You Might Also Like
Investor relations going sour?
Monday, April 29, 2024
We dive into the results of our latest survey, IBM goes to court and unexpected UK royalty in Dubai. View in browser ITA We make future final Good morning there, Wow, founders are going through a
📈 What’s getting funded now
Sunday, April 28, 2024
The Lightcone crew dives into the trends they're seeing from the latest YC batch. | This Week at YC April 28th, 2024 What a week! The deadline* for Summer 2024 applications was on Monday, which
Mailbob, Athina AI, Style AI, and Formula Wizard
Sunday, April 28, 2024
WebsiteGPT to make fully customizable websites in seconds BetaList BetaList Daily Mailbob Exclusive Perk publish and manage your email newsletter with ease Athina AI Exclusive Perk Monitor and Evaluate
Initiator Creator - Issue 140
Sunday, April 28, 2024
Initiator Creator - Issue #140 - ( Read in browser ) By Saurabh Y. // 28 April 2024 Presented by DesignThingy This Week's Notes: Directionless bull All our efforts get wasted when we get stuck
Knowing is not enough
Sunday, April 28, 2024
you must apply , The Start Your Ecommerce Business 2024 was a huge success thanks in part to you. We hope you left feeling inspired and equipped with actionable strategies from our expert speakers. But
#177 | Request for Startups, Real Estate Tech Themes, & more
Sunday, April 28, 2024
April 22nd–28th | Insights from Crossbeam, Altimeter, Ansa, Scale VC, and others ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
The $4M SaaS nobody knows about
Sunday, April 28, 2024
Starter Story's Sunday Breakfast #90 ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
Twitter’s former Head of Product opens up: being fired, meeting Elon, changing stagnant culture, building consumer…
Sunday, April 28, 2024
Kayvon was the longest-serving head of product at Twitter and was GM of Twitter's consumer division until the platform was acquired by Elon Musk. He originally joined Twitter in 2015 through the
Parea AI, Notion to Calendar, and VisualInvoicer
Saturday, April 27, 2024
Simplest way to create invoices BetaList BetaList Daily Parea AI Exclusive Perk Developer tools for evaluating, testing and monitoring LLM-powered apps. Notion to Calendar Put your Notion Databases in
From 0 experience to 160 orders for her business
Saturday, April 27, 2024
Read foundr+ student wins , Here at Foundr, we can't get enough of our student wins, both big and small. (These student wins are just the tip of the iceberg when you join foundr+.) Whether it was