@factory Caught your presentation at Gartner. Recommend you show the brownie first, then the recipe second. A system intelligent enough to act as a real-time model router is killer. Then show how you do it and why this is the future.
I used to drive. I loved driving. But now, I love thinking. Talking with my spouse. Looking at scenery. Chatting with Grok.
I think typing is following a similar arc. However, there are times I need to type to help my brain formulate the correct arc. This takes time. The edit/read/modify cycle is important (at times). @typelessdotcom is my daily keyboard driver, but not ideal in 100% of the use cases.
@itsdavidramos@hnshah@SamAsante At the pace of change, flat fee pricing models put the vendor at risk by marginalizing revenue over the long run. I want my tool vendors to survive for the long run. You’ll want my $2.95 in month 27 and I’ll want your updates if you sustain relevance.
@wadefoster@zapier It would be terrible to have one clear winner because that would signal the end of innovation and competitive pressure. An AI race that is never won, only led, is ideal.
Indeed. Coda, Notion, even Google Workspace cannot put a sizable dent in the requirements you are likely envisioning.
My take—discrete ontologies that can be federated and merged. Topic map specs support this. Palantir does it. A common enterprise platform like Palantir’s doesn’t exist as far as I know.
Ontological representation of everything is a big lift that is lighter with AI, but still non-trivial.
@Mahy__K@agupta Two things -
1. It does exist - Spark Desktop app is secretly a CLI. Claude can fully control it.
2. The value of solutions is not in the code. Email is deceptively complex. Requires a lot of “stuff”.
@trwpang_@agupta@t_blom Tom - are you on Spark Beta? They’ve made the desktop binary a CLI. I use it from Claude, Kiro, Antigravity, etc. they also have a beta for agentic rules.
@clarityx And, you will not see him leaving, coming.
In the vastly compressed AI timeline, I predict his tenure at Anthropic will mirror the mere milliseconds a stone touches a smooth lake on a single skip.
@getSortedAppios This was written before "we" realized agents need humans as much as humans need agents. @runpentagon is a new and interesting example.
An individual can now capture personal value extremely cheaply by cloning or approximating software via AI-assisted development. This demonstrates how subjective value can decouple from traditional production costs.
However, the market still heavily rewards the original product because:
• Most buyers (companies) value reliability, security, support, and network effects more than raw functionality.
• Building and sustaining what the market values (polish, scale, trust) remains hard and expensive, even if the core code is replicable.
• The original company can keep iterating faster than scattered clones.
In short: Individual value is about personal fit and “good enough.” Market value is about coordinated willingness to pay across many people, which favors superior total solutions over isolated clever hacks.
You created a clever hack. Good for you. Literally, for you. I often argue that non-consumption is the biggest and most silent competitor every product faces.
However, taking a position that your code hack invalidates the bundle of goods @linear has assembled, is foolish thinking. Products are not code.