Newsletter: Are we sure Forward Deployed Engineers are a good idea?
And one 1700 football field sized data center
Brixo latest
Remember when only engineers could update websites?
Marketing wanted to change a headline — they had to open a ticket.
Product wanted to tweak layout — buried in the release queue.
That bottleneck slowed everything down.
Then the tools changed — and everything opened up.
A quick timeline:
2000: Websites were hand-coded in HTML and deployed manually by engineers.
2000–2010: CMS platforms like WordPress and Drupal gave marketers limited control — they could publish, but not design.
2010–2015: Tools like Squarespace and Shopify made it easier for teams outside engineering to build and manage content.
2015–2020: Webflow and Framer made design and publishing truly visual — no dev tickets required.
2023+: “Vibe coding” and AI-assisted creation are letting anyone spin up sites, campaigns, and apps in real time.
Each shift made the web more collaborative, accessible, and fast.
Agents are at that same inflection point today.
Right now, only engineering can see what’s going on.
That won’t last.
What stood out
FDE’s are being glorified as the ultimate solution lately. While it may be the right solution today, it’s symbolic of the state of agent & AI technology today.
It’s clearly too difficult for even smart folks (see comment). This is why the parallel to marketing websites makes sense. We know it won’t be difficult for too long because it can’t. The domain experts will need visibility and control to ensure agents are successful.
What we read/listened/watched
Claude can now use Skills to improve how it performs specific tasks. Skills are folders that include instructions, scripts, and resources that Claude can load when needed.
We believe skills are step in the right direction in making agents easier to manage.
Your AI tools run on fracked gas and bulldozed Texas land
This article provides some great visual examples to help you understand just how big these data centers are going to be.
Watching NFL Sunday now and can’t believe 1 data center will be 1700 football fields!
Evals Are the New Moat—And RL Turns Them into Your Product Advantage
Eric Fett from NGP Capital argues that evals are the new moat. A worthwhile read if you’re trying to understand the landscape.
Final brick
The purpose of the Final Brick is to leave you with something to think about.
Jason Lemkin (yes featured twice) details a situation he experienced as a board member. It’s surprising for a few reasons:
$40M ARR is a big company but not so big that you live under a rock. It would be very difficult to not have heard of “this guy” Replit.
This is likely common. Most companies of this size live on zoom all-day, which limits their ability to experiment and learn new techniques.
It proves the reality - there is so much opportunity to out-execute with new AI tools.
If you’re not out there experimenting, there’s no better time to start than now.