Devolver

SignalWire is a client of mine. They shipped POM in April.

We noticed that Microsoft released the POML (Prompt Orchestration Markup Language) yesterday.

SignalWire's version complies to markdown or XML.

The very close parallels may be pure coincidence, but it may not be.

If a clone job...not a good look to rip off OSS software without even a link to the inspiration.

If a coincidence: the co-founders at SignalWire have a lot of foresight (they hacked a tool use framework within two weeks of GPT-4, before OpenAI introduced tool calling).

Their AI toolset is worth a deeper look.

Surprised not to see Bento on the consideration list. Maybe not (yet) well known, but gives Customer.io a run for its money and pricing is great.

A "ChatGPT app designed by Apple" scenario would be the safest, simplest, and lowest-risk partnership for both companies.

It would also be the least imaginative.

The wilder, more interesting scenarios:

- "SiriGPT" - GPT in Apple data centers, running on Apple Silicon - Why not both?

In all three, Apple partners with OpenAI to iterate Apple Silicon towards performance parity with NVIDIA...but with much better performance-per-watt.

Totally agree with you.

The essay is less about the regulation and more about why Google needs HubSpot more than HubSpot needs Google.

As generative AI matures, HubSpot’s “all-in-one” software for marketing, sales, and customer support has an opportunity to evolve into AI-powered flywheel that accelerates (and eventually automates) every aspect of a business’s relationship with its prospects and customers.

Given the mixed history of large acquisitions and the growing field of AI models that aren’t Google’s, HubSpot seems more likely to actualize its biggest opportunities if it does not accept tempting acquisition offers from Google (or anyone else).