Sunny

← All weeks

What I’ve been reading (W4, 2023)

20 Things I’ve Learned in my 20 Years as a Software Engineer · Using the wrong tool for the job

20 Things I’ve Learned in my 20 Years as a Software Engineer

  • Designing software is primarily a listening activity; we often have to be software engineers, part psychics, and part anthropologists. Investing in this design process through dedicated UX team members or by simply educating yourself will deliver enormous dividends.
  • Great software engineers think deeply about the user experience of their code — external API, programmatic API, user interface, protocol, or any other interface; great engineers consider who will be using it, why it will be used, how it will be used, and what is important to those users.
  • The best code is no code, or code you don’t have to maintain
  • The primary job of any software engineer is delivering value. Truly internalizing this leads to a different way of solving problems and viewing your tools.
  • Suppose you don’t understand what is possible or available in a given ecosystem. In that case, you’ll find it impossible to design a reasonable solution to all but the most simple problems. To summarize, be wary of people designing systems who haven’t written code in a long time.
  • Worry less about elegance and perfection; instead, strive for continuous improvement and create a livable system where your team enjoys working and sustainably delivers value.
  • One of the biggest differences between a most significant engineer and a junior engineer is that they’ve formed opinions about the way things should be
  • Dealing with this data in the future can become a nightmare. Just remember, your data will likely long outlive your codebase. Spend energy keeping it orderly and clean; it’ll pay off well in the long run. (my favourite cc: )

Using the wrong tool for the job

  • Impedance is when you have to take on complexity to solve a problem due to your choice of tools.
  • The more impedance, the more “wrong” your tool is for the job.
  • The complexity from impedance usually (usually) is counteracted by the higher complexity ceilings you get with familiar tools and possibly by the additional simplifications you get from tool mastery.
  • Impedance isn’t a one-and-done deal: as a system grows, you need to do more things, which adds more impedance (and more places for worse impedance).
  • The ecosystem growth also reduces the impedance of the “wrong tool”, so it’s more viable in place of the “right tool”.

How Amazon Uses AI To Automate Work In Its Corporate Headquarters

  • Amazon leverages the predictable nature of humans to automate tasks through the Hands of the Wheel program.
  • Senior leadership at Amazon started exploring AI potential in 2012, with Ralf Herbrich leading the effort to bring Project Yoda to life.
  • Herbrich’s team of machine-learning scientists spent several years improving the accuracy of AI predictions.
  • AI predictions were integrated into employees’ workflow tools and eventually replaced human decision-making in forecasting, pricing, purchasing, inventory planning, merchandising marketing, and negotiation with suppliers.

Others

  • We are so conditioned to think that our lives revolve around great moments. But great moments often catch us unawares: https://kentnerburn.com/the-cab-ride-ill-never-forget
  • I think if you can focus on the important things and not do the things that waste time, you can be fairly sloppy with productivity otherwise, and you’ll still get far more done than most people: https://youtu.be/iRwk9UajXFg?t=333
  • The machines replacing humans narrative is powerful. But the narrative I prefer is that AI is making things available that have been expensive and unobtainable for so many. And that is not limited to programming. It is true of so many things: https://avc.com/2023/01/the-ai-assist/
Substack edition

View the original post

20 Things I’ve Learned in my 20 Years as a Software Engineer

  • Designing software is primarily a listening activity; we often have to be software engineers, part psychics, and part anthropologists. Investing in this design process through dedicated UX team members or by simply educating yourself will deliver enormous dividends.

  • Great software engineers think deeply about the user experience of their code — external API, programmatic API, user interface, protocol, or any other interface; great engineers consider who will be using it, why it will be used, how it will be used, and what is important to those users.

  • The best code is no code, or code you don’t have to maintain

  • The primary job of any software engineer is delivering value. Truly internalizing this leads to a different way of solving problems and viewing your tools.

  • Suppose you don’t understand what is possible or available in a given ecosystem. In that case, you’ll find it impossible to design a reasonable solution to all but the most simple problems. To summarize, be wary of people designing systems who haven’t written code in a long time.

  • Worry less about elegance and perfection; instead, strive for continuous improvement and create a livable system in which your team enjoys working and sustainably delivers value.

  • One of the biggest differences between a most significant engineer and a junior engineer is that they’ve formed opinions about the way things should be

  • Dealing with this data in the future can become a nightmare. Just remember, your data will likely long outlive your codebase. Spend energy keeping it orderly and clean; it’ll pay off well in the long run.

Using the wrong tool for the job

  • Impedance is when you have to take on complexity to solve a problem due to your choice of tools.

  • The more impedance, the more “wrong” your tool is for the job.

  • The complexity from impedance usually (usually) is counteracted by the higher complexity ceilings you get with familiar tools and possibly by the additional simplifications you get from tool mastery.

  • Impedance isn’t a one-and-done deal: as a system grows, you need to do more things, which adds more impedance (and more places for worse impedance).

  • The ecosystem growth also reduces the impedance of the “wrong tool”, so it’s more viable in place of the “right tool”.

How Amazon Uses AI To Automate Work In Its Corporate Headquarters

  • Amazon leverages the predictable nature of humans to automate tasks through the Hands of the Wheel program.

  • Senior leadership at Amazon started exploring AI potential in 2012, with Ralf Herbrich leading the effort to bring Project Yoda to life.

  • Herbrich’s team of machine-learning scientists spent several years improving the accuracy of AI predictions.

  • AI predictions were integrated into employees’ workflow tools and eventually replaced human decision-making in forecasting, pricing, purchasing, inventory planning, merchandising marketing, and negotiation with suppliers.

Others

Thanks for reading Notes! Subscribe for free to receive new posts and support my work.

Originally on Substack