Ted Slocum's Weblog

Another interesting idea from this podcast is that agents that are shitty won’t cut it. If you just tack on agents that aren’t thoughtful and high quality to existing products / software, people won’t pay for them.

They were talking in the context of the PE strategy of acquiring a company and instituting the change of “we need an AI agent product” to add value to a company. But I think it applies at a much smaller scale too — SMB are not going to finance and pay for mediocre agents either. Whether they are custom agents or agents in existing ecosystems (like Copilot Studio), they will need to be great products.

60% agents won’t cut it

— Jason Lemkin

To help out with link blogging I created a few skills (technically they are /commands).

• /link-new to create a new post file with frontmatter pre-populated (link)

• /link-review to proof your post to make sure it is up to your standards (link)

• /link-publish to publish your post (link)

Bonus

• /photo to take a photo from your clipboard and host it for you to easily insert in your post (link)

“Automations” like these are cool because they break the stereotype that automations hinder creativity, thinking and collaboration — these ones push me to think more, be original, and share!

I was inspired by My approach to running a link blog.

I just found this podcast and agree with this commenter

Man this podcast is criminally underrated. Love the unadulterated straight shooting. Keep it up gents! — @howtoactuallyinvest

No politics or virtue signaling!

---

One interesting take was in the Anthropic vs OpenAI battle. One way to frame why Anthropic is rolling and OpenAI is reeling (relatively) is that consumers won’t pay for intelligence like enterprises will. When you’re at home you want netflix; when you’re at work you want intelligence. It could be that Anthropic’s decision to go after enterprise rather than consumer was literally a trillion dollar decision!

A few take-aways from this podcast. There were many more and I highly recommend watching

The employees are the company

This was in the context of building a company and recruiting. Take away for me, never make any “mehh” hires unless you want a mediocre company.

The idea of a PM makes no sense in the future

The process of taking requests, forming year + roadmaps for features is all getting compressed… Relatedly

All proposed features need to be demos — not slide shows.

And lastly, the skills that are going to be useful are in line with being a CEO

What are we building and why?

I was happy to find this tool by Simon Willison that solved a real annoyance: copy and pasting code from a terminal running Claude Code. The results are a mess. I found myself prompting claude to save the output to a file just so I didn’t have to reformat things manually.

I liked this tool so much, but it was still a hassle to use Simon’s web tool. So I turned the idea into a Raycast script. It reads the current content of the clipboard and pastes the cleaned up text with command + option + v