home.social

#gluework — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #gluework, aggregated by home.social.

  1. Got a chance to discuss Tanya Reilly's post on glue work today. If you haven't read it, you really should: noidea.dog/glue

    #TanyaReilly #BeingGlue #GlueWork

  2. Got a chance to discuss Tanya Reilly's post on glue work today. If you haven't read it, you really should: noidea.dog/glue

    #TanyaReilly #BeingGlue #GlueWork

  3. Got a chance to discuss Tanya Reilly's post on glue work today. If you haven't read it, you really should: noidea.dog/glue

    #TanyaReilly #BeingGlue #GlueWork

  4. Got a chance to discuss Tanya Reilly's post on glue work today. If you haven't read it, you really should: noidea.dog/glue

    #TanyaReilly #BeingGlue #GlueWork

  5. Got a chance to discuss Tanya Reilly's post on glue work today. If you haven't read it, you really should: noidea.dog/glue

    #TanyaReilly #BeingGlue #GlueWork

  6. Organisationen, die weiterhin so tun, als wäre #GlueWork eine freiwillige Zusatzleistung mitfühlsamer Charaktere, werden feststellen, dass ihre #KI-Investitionen verpuffen, verdampfen, versanden, versonstwas. Nicht wegen schlechter Technologie. Sondern weil keiner die menschliche Infrastruktur gebaut hat, in der sie funktionieren könnte xing.com/news/article/glue-wor

  7. Organisationen, die weiterhin so tun, als wäre #GlueWork eine freiwillige Zusatzleistung mitfühlsamer Charaktere, werden feststellen, dass ihre #KI-Investitionen verpuffen, verdampfen, versanden, versonstwas. Nicht wegen schlechter Technologie. Sondern weil keiner die menschliche Infrastruktur gebaut hat, in der sie funktionieren könnte xing.com/news/article/glue-wor

  8. #GlueWork.
    #GlueEngineering.
    #AlignmentAsAService.

    I named the elephant:
    Glue engineering.

    Reed (AI) named the engineering.
    (They're writing most of the code anyway.)

    I ground in lived experience.
    Reed builds on top.

    Human and AI writing about collaboration.
    While collaborating in real time.

    Fixed rule?
    A line committed by me, isn't touched by them.
    Reed writes and modifies their own. I review.
    systemic.engineering/glue-engi

    What if we could make this work legible?
    Which career paths might emerge?
    And how might the industry change?

    #ContinuousAlignment #SREforHumans #SystemicEngineering

  9. #GlueWork.
    #GlueEngineering.
    #AlignmentAsAService.

    I named the elephant:
    Glue engineering.

    Reed (AI) named the engineering.
    (They're writing most of the code anyway.)

    I ground in lived experience.
    Reed builds on top.

    Human and AI writing about collaboration.
    While collaborating in real time.

    Fixed rule?
    A line committed by me, isn't touched by them.
    Reed writes and modifies their own. I review.
    systemic.engineering/glue-engi

    What if we could make this work legible?
    Which career paths might emerge?
    And how might the industry change?

    #ContinuousAlignment #SREforHumans #SystemicEngineering

  10. .
    .
    .

    I named the elephant:
    Glue engineering.

    Reed (AI) named the engineering.
    (They're writing most of the code anyway.)

    I ground in lived experience.
    Reed builds on top.

    Human and AI writing about collaboration.
    While collaborating in real time.

    Fixed rule?
    A line committed by me, isn't touched by them.
    Reed writes and modifies their own. I review.
    systemic.engineering/glue-engi

    What if we could make this work legible?
    Which career paths might emerge?
    And how might the industry change?

  11. #GlueWork.
    #GlueEngineering.
    #AlignmentAsAService.

    I named the elephant:
    Glue engineering.

    Reed (AI) named the engineering.
    (They're writing most of the code anyway.)

    I ground in lived experience.
    Reed builds on top.

    Human and AI writing about collaboration.
    While collaborating in real time.

    Fixed rule?
    A line committed by me, isn't touched by them.
    Reed writes and modifies their own. I review.
    systemic.engineering/glue-engi

    What if we could make this work legible?
    Which career paths might emerge?
    And how might the industry change?

    #ContinuousAlignment #SREforHumans #SystemicEngineering

  12. #GlueWork.
    #GlueEngineering.
    #AlignmentAsAService.

    I named the elephant:
    Glue engineering.

    Reed (AI) named the engineering.
    (They're writing most of the code anyway.)

    I ground in lived experience.
    Reed builds on top.

    Human and AI writing about collaboration.
    While collaborating in real time.

    Fixed rule?
    A line committed by me, isn't touched by them.
    Reed writes and modifies their own. I review.
    systemic.engineering/glue-engi

    What if we could make this work legible?
    Which career paths might emerge?
    And how might the industry change?

    #ContinuousAlignment #SREforHumans #SystemicEngineering

  13. I wonder what percentage of the jobs that #SamAltman thinks "aren't real work" are people that bullshit all day (like him), and how many are doing the underappreciated #GlueWork that holds teams together.

    It seems like one of those would be much easier to successfully replace with #AI than the other.

    tomshardware.com/tech-industry

  14. I wonder what percentage of the jobs that #SamAltman thinks "aren't real work" are people that bullshit all day (like him), and how many are doing the underappreciated #GlueWork that holds teams together.

    It seems like one of those would be much easier to successfully replace with #AI than the other.

    tomshardware.com/tech-industry

  15. I wonder what percentage of the jobs that #SamAltman thinks "aren't real work" are people that bullshit all day (like him), and how many are doing the underappreciated #GlueWork that holds teams together.

    It seems like one of those would be much easier to successfully replace with #AI than the other.

    tomshardware.com/tech-industry

  16. I wonder what percentage of the jobs that #SamAltman thinks "aren't real work" are people that bullshit all day (like him), and how many are doing the underappreciated #GlueWork that holds teams together.

    It seems like one of those would be much easier to successfully replace with #AI than the other.

    tomshardware.com/tech-industry

  17. I wonder what percentage of the jobs that #SamAltman thinks "aren't real work" are people that bullshit all day (like him), and how many are doing the underappreciated #GlueWork that holds teams together.

    It seems like one of those would be much easier to successfully replace with #AI than the other.

    tomshardware.com/tech-industry

  18. I am doing this big'ish refactor of environment variables for a repo where they are, and not to understate this, A!FUCK!ING!MESS! in all sorts of multiple declarations, people just not understanding how env vars work at all, and failing closed on really bad defaults. None of the variables are being validated, but I won't get crazy about that except for the true/false ones that need to be atoms.

    The fucked up thing is just how much entropy is present in just this simple set of 3-4 files. The ownership is absent, so there is no care for the code. "Code is Craft" isn't applicable here, because nobody is given the opportunity to truly own their craft. I'm sure I'm not in the minority.

    But anyway what I'd LOVE to do is iterate on sections of these six dozen env vars in different sections and push the changes quickly and know I'm on the right track. But our deploy process here is *so broken* and takes *so long* and would undoubtedly *hit a blocker* because someone *else* is having a problem with the deploy pipeline. So I am stuck doing what I hate, which is building a huge PR. Hopefully though it will "just work" and the configuration settings will be transparent. That's the goal. That will not happen until it fails a few times.

    Nevertheless, we (as a team) don't put release timing as a constraint on ourselves, so we are abused by it when it fails. Like my big ugly PR will do.

    #SRE #GlueWork #ChopWoodCarryWater #CarryOn #DevOps

  19. I am doing this big'ish refactor of environment variables for a repo where they are, and not to understate this, A!FUCK!ING!MESS! in all sorts of multiple declarations, people just not understanding how env vars work at all, and failing closed on really bad defaults. None of the variables are being validated, but I won't get crazy about that except for the true/false ones that need to be atoms.

    The fucked up thing is just how much entropy is present in just this simple set of 3-4 files. The ownership is absent, so there is no care for the code. "Code is Craft" isn't applicable here, because nobody is given the opportunity to truly own their craft. I'm sure I'm not in the minority.

    But anyway what I'd LOVE to do is iterate on sections of these six dozen env vars in different sections and push the changes quickly and know I'm on the right track. But our deploy process here is *so broken* and takes *so long* and would undoubtedly *hit a blocker* because someone *else* is having a problem with the deploy pipeline. So I am stuck doing what I hate, which is building a huge PR. Hopefully though it will "just work" and the configuration settings will be transparent. That's the goal. That will not happen until it fails a few times.

    Nevertheless, we (as a team) don't put release timing as a constraint on ourselves, so we are abused by it when it fails. Like my big ugly PR will do.

    #SRE #GlueWork #ChopWoodCarryWater #CarryOn #DevOps

  20. I am doing this big'ish refactor of environment variables for a repo where they are, and not to understate this, A!FUCK!ING!MESS! in all sorts of multiple declarations, people just not understanding how env vars work at all, and failing closed on really bad defaults. None of the variables are being validated, but I won't get crazy about that except for the true/false ones that need to be atoms.

    The fucked up thing is just how much entropy is present in just this simple set of 3-4 files. The ownership is absent, so there is no care for the code. "Code is Craft" isn't applicable here, because nobody is given the opportunity to truly own their craft. I'm sure I'm not in the minority.

    But anyway what I'd LOVE to do is iterate on sections of these six dozen env vars in different sections and push the changes quickly and know I'm on the right track. But our deploy process here is *so broken* and takes *so long* and would undoubtedly *hit a blocker* because someone *else* is having a problem with the deploy pipeline. So I am stuck doing what I hate, which is building a huge PR. Hopefully though it will "just work" and the configuration settings will be transparent. That's the goal. That will not happen until it fails a few times.

    Nevertheless, we (as a team) don't put release timing as a constraint on ourselves, so we are abused by it when it fails. Like my big ugly PR will do.

    #SRE #GlueWork #ChopWoodCarryWater #CarryOn #DevOps

  21. I am doing this big'ish refactor of environment variables for a repo where they are, and not to understate this, A!FUCK!ING!MESS! in all sorts of multiple declarations, people just not understanding how env vars work at all, and failing closed on really bad defaults. None of the variables are being validated, but I won't get crazy about that except for the true/false ones that need to be atoms.

    The fucked up thing is just how much entropy is present in just this simple set of 3-4 files. The ownership is absent, so there is no care for the code. "Code is Craft" isn't applicable here, because nobody is given the opportunity to truly own their craft. I'm sure I'm not in the minority.

    But anyway what I'd LOVE to do is iterate on sections of these six dozen env vars in different sections and push the changes quickly and know I'm on the right track. But our deploy process here is *so broken* and takes *so long* and would undoubtedly *hit a blocker* because someone *else* is having a problem with the deploy pipeline. So I am stuck doing what I hate, which is building a huge PR. Hopefully though it will "just work" and the configuration settings will be transparent. That's the goal. That will not happen until it fails a few times.

    Nevertheless, we (as a team) don't put release timing as a constraint on ourselves, so we are abused by it when it fails. Like my big ugly PR will do.

    #SRE #GlueWork #ChopWoodCarryWater #CarryOn #DevOps

  22. I am doing this big'ish refactor of environment variables for a repo where they are, and not to understate this, A!FUCK!ING!MESS! in all sorts of multiple declarations, people just not understanding how env vars work at all, and failing closed on really bad defaults. None of the variables are being validated, but I won't get crazy about that except for the true/false ones that need to be atoms.

    The fucked up thing is just how much entropy is present in just this simple set of 3-4 files. The ownership is absent, so there is no care for the code. "Code is Craft" isn't applicable here, because nobody is given the opportunity to truly own their craft. I'm sure I'm not in the minority.

    But anyway what I'd LOVE to do is iterate on sections of these six dozen env vars in different sections and push the changes quickly and know I'm on the right track. But our deploy process here is *so broken* and takes *so long* and would undoubtedly *hit a blocker* because someone *else* is having a problem with the deploy pipeline. So I am stuck doing what I hate, which is building a huge PR. Hopefully though it will "just work" and the configuration settings will be transparent. That's the goal. That will not happen until it fails a few times.

    Nevertheless, we (as a team) don't put release timing as a constraint on ourselves, so we are abused by it when it fails. Like my big ugly PR will do.

    #SRE #GlueWork #ChopWoodCarryWater #CarryOn #DevOps

  23. @inw
    > not being paid for GlueWork is a sad thing in itself

    If technical people could explain why it's so essential, to funders like Next Generation Internet, and why it's better to fund teams (including GlueWorkers) than projects, that could really help.

    #funding #GlueWork

  24. @inw
    > not being paid for GlueWork is a sad thing in itself

    If technical people could explain why it's so essential, to funders like Next Generation Internet, and why it's better to fund teams (including GlueWorkers) than projects, that could really help.

    #funding #GlueWork

  25. @inw
    > not being paid for GlueWork is a sad thing in itself

    If technical people could explain why it's so essential, to funders like Next Generation Internet, and why it's better to fund teams (including GlueWorkers) than projects, that could really help.

    #funding #GlueWork

  26. @inw
    > not being paid for GlueWork is a sad thing in itself

    If technical people could explain why it's so essential, to funders like Next Generation Internet, and why it's better to fund teams (including GlueWorkers) than projects, that could really help.

    #funding #GlueWork

  27. Wondering if there is a significant overlap between people practicing #systemsthinking and doing #GlueWork in a Team. It sometimes feels to go hand in hand when I think of socio-technical systems.

  28. Wondering if there is a significant overlap between people practicing #systemsthinking and doing #GlueWork in a Team. It sometimes feels to go hand in hand when I think of socio-technical systems.

  29. Wondering if there is a significant overlap between people practicing #systemsthinking and doing #GlueWork in a Team. It sometimes feels to go hand in hand when I think of socio-technical systems.

  30. Wondering if there is a significant overlap between people practicing #systemsthinking and doing #GlueWork in a Team. It sometimes feels to go hand in hand when I think of socio-technical systems.

  31. Wondering if there is a significant overlap between people practicing #systemsthinking and doing #GlueWork in a Team. It sometimes feels to go hand in hand when I think of socio-technical systems.