What’s 🔥 in Enterprise IT/VC #460
Bubble 🫧 or just a bump in the road?
I’m keeping this week’s commentary short as the links below tell the story.
The market is feeling frothy. Inception and seed rounds are hitting record sizes. Investors are arguing whether negative gross margin businesses can ever be true PMF. Founders are cashing out big in secondaries. And now, thin-sliced reports claiming “AI isn’t delivering ROI” are rattling already-nervous markets.
That being said, we are busier than ever, having signed one inception stage term sheet this week and hoping to close out another one this weekend. As we head into the final weeks of summer, I just want to remind everyone that we are still just in the early innings. Even if the foundational models don’t improve much from here, there is still so much we can all automate with existing technology that will take over a decade to deliver.
Channel your inner Warren Buffet and remember:
“Be fearful when others are greedy, and be greedy when others are fearful.”
As always, 🙏🏼 for reading and please share with your friends and colleagues!
Scaling Startups
#💯 not everything is 📈 from inception
#some great data for founders from Peter Walker, Carta (full deck 50 slides here)
doing more with less…
#🤔 must lead with 10x better product at a minimum
Enterprise Tech
#🫧 bubble?
#sign of times as well as VCs keep investing and buying them out
#oh no the sky is falling - MIT drops study on AI not working in enterprises and market cites that as huge reason for stock drops but we all know it’s still so early!
#AI house of cards?
#because VCs continue to have FOMO and pour 💰 into startups forgetting ZIRP era
#to that end 🎯 👇🏼 - also this is 🔑 and something a founder in automation space told me back in 2018 - when selling to enterprises he would always lead with, “we can help you understand your processes but automating shitty processes is not what is best, you have to be willing to reimagine them!”
#AI and agents seem pretty effective to me…
#voice AI so underrated and a simple but amazing use case for voice AI
#highly worth reading…even South Park Episode 3 had some references to semiconscious AI
#more deep thinking 🤔
#🎯
#must read, how to use AI to code and teach it to continuously learn and make it more deterministic?
#evals, evals, evals - repeat after me…
#simulations are the future - here’s what Sierra is doing but lots of new startups being created to solve this problem…
Agents need a totally different approach because with AI the same inputs will produce different outputs — making it hard to know if an agent is “working,” let alone production-ready. So the key question is not whether an agent did what it was told but whether it enables customers to accomplish their goals.
Simulated conversations between agents and mock personas ensure reliability at scale. They are a new kind of testing, for a new kind of software.
#speaking of simulations, how about simulating Microsoft purely with AI?
#this is 🤯 - build your own simulation of worlds…live - any to the point above can use to train robotics, prepare for disasters and much more…
#on Cursor, subsidies and negative gross margin - great post from Chris Paik
#absolutely huge - memory for specific projects
#no matter how much you spend on security, humans are oftentimes the weakest link
#such a good question and click 👇🏻 for comments
#this is very good 👇🏼
#the edge is the future, will be a massive battleground as individuals want privacy, where speed matters as devices get better chips to do on device processing…apple had the right to win but…
Markets
#🤯 - we are now in Series K land! What happens after Z?
#Netskope files for IPO! (Bloomberg)
Netskope Inc., a cloud security provider, filed for a US initial public offering, disclosing growing revenue and shrinking losses.
The Santa Clara, California-based company had net loss of $169.5 million on revenue of $328.5 million for the six months ended July 31, compared with net loss of $206.7 million on revenue of $251.3 million in the corresponding period a year earlier, according to a filing with the US Securities and Exchange Commission on Friday.
Founded in 2012, the company’s cloud-native Netskope One platform incorporates artificial intelligence models to help companies protect customers from threats and keep sensitive data secure, according to the filing.