4 ms·

> One approach would be to monitor what management considers productive / successful employees and also on employees who have gone on PIP's - then train the AI

by er4hn 5y ago

> One approach would be to monitor what management considers productive / successful employees and also on employees who have gone on PIP's - then train the AI on that data set. I look forward to the papers on how to slack off on your job while the AI classifies you as "a straight shooter with upper management written all over him". Possible title: "An adversarial approach to the TwoBobs Productivity Classifier" (/s if this was not clear)

Terretta 5y ago
Sadly, not /s. I’ve been pitched the TwoBobs Productivity Classifier. My feedback was anyone who needs it has classified themselves as an unproductive dev manager. I suggested the concept be made available as being by devs, for devs, purposed as a personal coach, with no uplink to the mothership.
- monsieurbanana 5y ago
  > I suggested the concept be made available as being by devs, for devs, purposed as a personal coach, with no uplink to the mothership. Ah yes, the classic "foot in the door" approach. Despite all the good intentions behind it, it still would end badly.
  - DaiPlusPlus 5y ago
    As long as there's more demand than supply for SWE workers, something like that would never become the norm. All of the dystopic news stories of WFH employees being subject to monitoring are invariably from fungible employees.
    - vbezhenar 5y ago
      I often read on HN that it's a norm for US employers to install spying software on developers computers. I don't see how is it different. I would try to avoid from working in such a companies, because I don't feel secure knowing that someone is spying on me.
      - DaiPlusPlus 5y ago
        > I often read on HN that it's a norm for US employers to install spying software on developers computers. I haven't heard of this happening at all. Whereas it is commonplace for people to use time-tracking software (like Toggl, RescueTime, etc) - and it is also not-uncommon for companies to require employees to use such software (software-eng or otherwise) for the purposes of reporting their time/hours/what-they're-doing, they all have web-based UIs for doing data-entry: which is definitely not the same thing as installing monitoring software on an employees' own personal hardware with or without their informed consent and it being a condition of employment. I think people are getting terms-conflated and using overloaded terms with negative connotations: "time tracking software" which does active-software-application-window-tracking for the benefit of the direct user can easily be mischaracterized as "monitoring software".
        gailbear 5y ago
        Not only have I heard of this happening, it's literally on my computer right now! Most of the monitoring on it is automated - but I also write that software. There's not a lot of ML involved in the kinds of automated activity checking that my employer is doing, but if they suspected I was not being productive, they could basically pull my entire body of network requests, which at the very least has addresses of sites, and can tell you roughly how long someone spent on them. Since 90% of my work is on the internet, this is pretty effective spying. Interestingly, I'd much prefer Microsoft's algorithms classifying me as productive or unproductive, based on how I code - while there's a lot to be said for different patterns of work being equally productive, I trust Microsoft to implement the same heuristics across the board, and not take a look at my gender and race and make assumptions about me based on that. (Of course, since the ML data of productivity will come from nearly entirely white males, if it's only available to github enterprise cloud subscribers, then the ML data will be pretty biased as it is. Guess discrimination is inevitable!)

by er4hn 5y ago

Terretta 5y ago

Sadly, not /s. I’ve been pitched the TwoBobs Productivity Classifier. My feedback was anyone who needs it has classified themselves as an unproductive dev manager. I suggested the concept be made available as being by devs, for devs, purposed as a personal coach, with no uplink to the mothership.

monsieurbanana 5y ago
> I suggested the concept be made available as being by devs, for devs, purposed as a personal coach, with no uplink to the mothership. Ah yes, the classic "foot in the door" approach. Despite all the good intentions behind it, it still would end badly.
- DaiPlusPlus 5y ago
  As long as there's more demand than supply for SWE workers, something like that would never become the norm. All of the dystopic news stories of WFH employees being subject to monitoring are invariably from fungible employees.
  - vbezhenar 5y ago
    I often read on HN that it's a norm for US employers to install spying software on developers computers. I don't see how is it different. I would try to avoid from working in such a companies, because I don't feel secure knowing that someone is spying on me.
    - DaiPlusPlus 5y ago
      > I often read on HN that it's a norm for US employers to install spying software on developers computers. I haven't heard of this happening at all. Whereas it is commonplace for people to use time-tracking software (like Toggl, RescueTime, etc) - and it is also not-uncommon for companies to require employees to use such software (software-eng or otherwise) for the purposes of reporting their time/hours/what-they're-doing, they all have web-based UIs for doing data-entry: which is definitely not the same thing as installing monitoring software on an employees' own personal hardware with or without their informed consent and it being a condition of employment. I think people are getting terms-conflated and using overloaded terms with negative connotations: "time tracking software" which does active-software-application-window-tracking for the benefit of the direct user can easily be mischaracterized as "monitoring software".
      - gailbear 5y ago
        Not only have I heard of this happening, it's literally on my computer right now! Most of the monitoring on it is automated - but I also write that software. There's not a lot of ML involved in the kinds of automated activity checking that my employer is doing, but if they suspected I was not being productive, they could basically pull my entire body of network requests, which at the very least has addresses of sites, and can tell you roughly how long someone spent on them. Since 90% of my work is on the internet, this is pretty effective spying. Interestingly, I'd much prefer Microsoft's algorithms classifying me as productive or unproductive, based on how I code - while there's a lot to be said for different patterns of work being equally productive, I trust Microsoft to implement the same heuristics across the board, and not take a look at my gender and race and make assumptions about me based on that. (Of course, since the ML data of productivity will come from nearly entirely white males, if it's only available to github enterprise cloud subscribers, then the ML data will be pretty biased as it is. Guess discrimination is inevitable!)