7 Comments
User's avatar
static's avatar

Twilio is the one that doesn't make sense to me- don't all of these new vibe-apps need to send email/sms? Or do they only communicate via agent-native protocols?

On the other hand Canva and Figma are tools for a legacy form of development- they should go away. Figma in particular has always been the source of awkward handoffs- it's a drawing program more than it is a UX tool.

Vercel is just a terrible company, built on the back of the most overcomplicated Javascript/Typescript mess, it will collapse under the weight of its own mess.

I could have Monday.com cloned by next Tuesday.

Hubspot has allowed people to create their own huge data messes in there- I think that will stick around.

--

No Ads.

Mitchell Kosowski's avatar

The freight section is the sleeper insight here. Everyone's focused on the AI vs. SaaS narrative, but a 49% flatbed rejection rate (higher than the 2022 pandemic surge!) is a genuinely surprising signal about the physical economy

It makes me wonder if there's a connection to the AI buildout itself: all those data centers and GPU clusters need steel, concrete, and heavy equipment to get built.

Charlie Simionescu-Marin's avatar

With 60% of companies increasing tech spend and investing in AI, we should be asking what those companies actually need from AI vendors once the novelty wears off and they're trying to connect all of it to their existing data and operations. Recently wrote about that here __https://cwssm.substack.com/p/the-sentient-company__

toolina's avatar

This post really captures how AI is reshaping things, even in unexpected ways. Speaking of AI, I've been using an ai image describer (https://imagedescriber.dev) to help with content creation—it's handy for generating descriptions quickly. Thanks for the insights!

TRADE CRAFTERS's avatar

What’s striking here is how the story refuses to settle into something clean. On one side, AI is absorbing more of the incremental spend, quietly becoming the default destination for new dollars. On the other, parts of the old stack aren’t collapsing, they’re adapting, holding ground, even growing in pockets .

That kind of environment tends to confuse people because it doesn’t reward simple narratives. It rewards selectivity. The market isn’t killing software, it’s sorting it. And historically, those are the periods where dispersion widens and the average participant feels like nothing is working, while a handful of names quietly compound.

The same tension shows up in pricing. Earnings drift higher, multiples compress, capital gets more demanding. That’s usually the market asking a different question. Not whether growth exists, but whether it’s worth what people have already been paying for it.

Dan McRae's avatar

interesting observations on individual spending on AI.

David Mensah's avatar

The trucks data and rejection of transportation instructions is interesting. Have the number of tracks increased, decreased or remained the same for the assessed period ?