Most actions are visual. Chat in its existing form has a certain mental model of text to text or voice to voice .
An extended variation, which may be voice in / visual + voice out. Dynamic and generative UI and that looks nothing like chat in my sense is the end state.
Can be loosely called chat but would look and feel completely diffeeent
Also when most work is done under the radar by agents , job tracing and job tracking may be the larger visualization and communication objective
@GurutejRudra@marclou That’s the assumption, everyone doesn’t want only to be cheap. They want something to make their life better at a value they place on it
this is such an exciting time to be building. There are so many fundamental UX unresolved points
- like, what the operating surface would look like for most people
- How voice would blend into the core in a natural way experience
- whether social work would truly reside in existing tools like slack, or would an entirely new construct evolve
Everyone has the opportunity to define, in their own small way , how these evolve. And that is an exciting space to be in
@faqdev One with expertise baked in (not just workflows), proactive, built by people with relationships, distribution and data. Incumbents with data would always have a hook
@icreatelife Writing for people. It is incredibly frustrating to read something written by AI.
Corrections and cleanup is fine. Just not the primary thought and narrative
For most users, I believe a small subset of the email capabilities would probably be sufficient to do the job.
That said - would users really understand their delivery rates and know what to do when they get flagged as spam due to an unintentional issue? that may be the true test here
@bentossell - One thread for single threaded devices (AI in your TVs, Alexas etc) or casual tasks (managing my personal life)
- Multiple thread for more complex tasks (Engineering team, marketing teams, and in general work settings)
Agree it would be terribly inefficient to generate a new UX every time.
Though, once a component is generated, and if it ends up being a frequent use case : it may just harden itself into becoming a higher order component.
With MCP's (and without an aligned visual surface), everyone is already having their own visual representations. E.g. Thariq's chrome plugin earlier today was an interesting example of that -https://t.co/dMBuiQX8Tn
Email has been resilient through a sea of changes over the last three decades - chat, newsletter tab in gmail : and has still survived.
This time, your point makes sense. Email is nothing but a protocol, no different from an MCP, with a lot of prior baggage and issues we could do without. It’s hard to see it play out for long.
If content is going to be consumed from specific walled gardens - Twitter, LinkedIn alternatives - I guess some avatar of what Thariq put out today seems like a legitimate first step: https://t.co/dMBuiQWB3P
Who controls the sort and visibility there, and what happens when 99.9% of content is AI generated? Lets slop is a temporary problem with a resolution in the short-medium term
@b4dgerr When the metric is to have 0 bugs, that becomes someone’s biggest goals
That is definitely not the right lens. But in large companies have seen that is how targets get set
@trq212 Love the new real time voice mode, however - it isnt able to write code in that voice mode. any way we can have that added in, would be a key step in hands-free coding end-end