5 ms·

Terminal emulator requires SSO login and transfers telemetry associated to email

22 points by Mo3 5y ago

xg15 5y ago
> Finally, as a general UX improvement, it might be nice to not require GH auth out of the box, but after a few hours of usage, say "Consider logging in with GH to get ". That's a much less confronting way to present authentication. Nag screens in my terminal! I can't wait...
rvz 5y ago
But it is written in Rust though. /s I don't know why a Terminal needs a SSO login wall. Seems like after all the glory and hype that the product is getting from HN, the true reality of it is that it is a data collection farm. I don't think anyone wants that.
- Nextgrid 5y ago
  Because it's a "growth & engagement" cancer that happens to have some terminal-like functionality with the end-game being to use said growth/engagement to solicit unreasonable amounts of VC funding, not an actual terminal for the purpose to help you get work done.
detaro 5y ago
was a major topic in the main discussion of the product yesterday: https://news.ycombinator.com/item?id=30921231 https://news.ycombinator.com/item?id=30921231
danpalmer 5y ago
This is receiving a lot of criticism that feels somewhat naive. A terminal emulator with advanced features that are company/team aware and with collaboration features isn't a bad idea. Product telemetry to understand in general terms how your software is being used so that you can focus efforts isn't a bad idea. The problems with Warp are: 1. They are being vague about what telemetry they send, creating a risk that sensitive data flowing through the application may be included. 2. It can't be turned off, even temporarily. Under many computer security policies at companies this is going to prevent Warp from being used at all. My suggestion would be: - Remove the login requirement to use a fully local version of the application. - Separate telemetry from service-backed smart features, allow each to be toggled individually. - Make it clear that terminal text input/output is never shared, or be very specific about when it is so that users can understand the security risk. - Don't associate telemetry with users or anonymise where possible.
- Mo3 5y ago
  I'm unsure how you rationalize calling the criticism naive, while at the same time 5 minutes later acknowledging and suggesting solutions for the exact same.
  - danpalmer 5y ago
    The criticism seems to be very anti-telemetry, or completely rejecting the idea of the smart/collaboration features. I think the issues are more nuanced than that.
  - lcnPylGDnU4H9OF 5y ago
    Taken more generously: * There is a lot of criticism. * At least some of that criticism is bad in this particular way. (As per an individual opinion.) * Here's a way to communicate the same criticism that is not bad in that particular way. As compared to the bug report: > I downloaded warp for the first time based on a link someone sent me about it being in public beta, but I'm immediately presented with a login form that I cannot dismiss. I know I cannot be the only developer who would never do a cloud login to use a terminal, let alone do it without any indication why it's needed or what it implies will be collected/transmitted to the warp team. Which is describing the software working as the developers intended and calling that a bug. Criticism can be expected to be better.
- zachlloyd 5y ago
  Thanks for the feedback - I'm the founder at Warp. There was a ton of discussion on this yesterday with our ShowHN [0] and we are considering what changes to make based on the feedback (some of which are inline with your suggestions). We do publish all the telemetry we collect here: https://docs.warp.dev/getting-started/privacy#exhaustive-telemetry-table https://docs.warp.dev/getting-started/privacy#exhaustive-tel... And we are committed to making telemetry optional when we leave beta. And it's definitely the case that command input / output are never sent to the server. Login is a bit trickier as it's hard to build an app that leverages the internet and facilitates teamwork without understanding user identity. But point taken. [0] https://news.ycombinator.com/item?id=30921231 https://news.ycombinator.com/item?id=30921231
theodric 5y ago
I hesitated to install it because of the telemetry reported by other HN users. I eventually installed it because of curiosity, but blocked its connections, and it refused to work, so I relented and unblocked them because of curiosity. I deleted it because it was crap. iTerm does what I need and want. Warp does not. Tempest in a teacup!
Shadonototra 5y ago
not even microsoft tried that to do that, these devs are out of touch
Longlius 5y ago
As a general rule, terminal emulators are a class of software I don't trust without access to source code.

22 points by Mo3 5y ago

xg15 5y ago

> Finally, as a general UX improvement, it might be nice to not require GH auth out of the box, but after a few hours of usage, say "Consider logging in with GH to get ". That's a much less confronting way to present authentication. Nag screens in my terminal! I can't wait...

rvz 5y ago

But it is written in Rust though. /s I don't know why a Terminal needs a SSO login wall. Seems like after all the glory and hype that the product is getting from HN, the true reality of it is that it is a data collection farm. I don't think anyone wants that.

Nextgrid 5y ago
Because it's a "growth & engagement" cancer that happens to have some terminal-like functionality with the end-game being to use said growth/engagement to solicit unreasonable amounts of VC funding, not an actual terminal for the purpose to help you get work done.

detaro 5y ago

was a major topic in the main discussion of the product yesterday: https://news.ycombinator.com/item?id=30921231 https://news.ycombinator.com/item?id=30921231

danpalmer 5y ago

This is receiving a lot of criticism that feels somewhat naive. A terminal emulator with advanced features that are company/team aware and with collaboration features isn't a bad idea. Product telemetry to understand in general terms how your software is being used so that you can focus efforts isn't a bad idea. The problems with Warp are: 1. They are being vague about what telemetry they send, creating a risk that sensitive data flowing through the application may be included. 2. It can't be turned off, even temporarily. Under many computer security policies at companies this is going to prevent Warp from being used at all. My suggestion would be: - Remove the login requirement to use a fully local version of the application. - Separate telemetry from service-backed smart features, allow each to be toggled individually. - Make it clear that terminal text input/output is never shared, or be very specific about when it is so that users can understand the security risk. - Don't associate telemetry with users or anonymise where possible.

Mo3 5y ago
I'm unsure how you rationalize calling the criticism naive, while at the same time 5 minutes later acknowledging and suggesting solutions for the exact same.
- danpalmer 5y ago
  The criticism seems to be very anti-telemetry, or completely rejecting the idea of the smart/collaboration features. I think the issues are more nuanced than that.
- lcnPylGDnU4H9OF 5y ago
  Taken more generously: * There is a lot of criticism. * At least some of that criticism is bad in this particular way. (As per an individual opinion.) * Here's a way to communicate the same criticism that is not bad in that particular way. As compared to the bug report: > I downloaded warp for the first time based on a link someone sent me about it being in public beta, but I'm immediately presented with a login form that I cannot dismiss. I know I cannot be the only developer who would never do a cloud login to use a terminal, let alone do it without any indication why it's needed or what it implies will be collected/transmitted to the warp team. Which is describing the software working as the developers intended and calling that a bug. Criticism can be expected to be better.
zachlloyd 5y ago
Thanks for the feedback - I'm the founder at Warp. There was a ton of discussion on this yesterday with our ShowHN [0] and we are considering what changes to make based on the feedback (some of which are inline with your suggestions). We do publish all the telemetry we collect here: https://docs.warp.dev/getting-started/privacy#exhaustive-telemetry-table https://docs.warp.dev/getting-started/privacy#exhaustive-tel... And we are committed to making telemetry optional when we leave beta. And it's definitely the case that command input / output are never sent to the server. Login is a bit trickier as it's hard to build an app that leverages the internet and facilitates teamwork without understanding user identity. But point taken. [0] https://news.ycombinator.com/item?id=30921231 https://news.ycombinator.com/item?id=30921231

theodric 5y ago

I hesitated to install it because of the telemetry reported by other HN users. I eventually installed it because of curiosity, but blocked its connections, and it refused to work, so I relented and unblocked them because of curiosity. I deleted it because it was crap. iTerm does what I need and want. Warp does not. Tempest in a teacup!

Shadonototra 5y ago

not even microsoft tried that to do that, these devs are out of touch

Longlius 5y ago

As a general rule, terminal emulators are a class of software I don't trust without access to source code.