5 ms·

Author here - I'm serious when I say that I'm happy to be a sounding board if you're at a career crossroads. Twitter DMs are open @forrestbrazeal.

NikolaNovak 8y ago
I've been in IT (mostly the "Legacy" kind - Oracle/PeopleSoft ERP) for about 20 years now, and I fully empathize with the article. I've subconsciously, then consciously, had to keep "moving up" on the value scale, partially due to outsourcing and partially due to automation. From sysadmin type to infra architect, but then made a curious curve that still catches me by surprise into tech lead and now find myself in basically management. I struggle to understand what will give me better skills and career for future: developing people and org knowledge and experience, as I do in my current role, which just feels generic and not particularly useful to somebody who used to gobble technical manuals for fun (but apparently market finds value in technically knowledgeable people who can manage and communicate); or go back to my love of technical and get back hands-on... understanding it's an uphill battle at my age and may be a losing one with the amount of automation and "AI being around the corner (forever;)".
- arethuza 8y ago
  If you have a decent understand of ERP systems and projects it might be easier and more productive to transfer those skills to more modern cloud-hosted ERP products than go back to more generic development skills?
  - NikolaNovak 8y ago
    Thx; Agree, though I find many "cloud-hosted ERP products" are actually very traditional ERP products, just hosted by vendor (I remember the "Lightbulb" moment when I realized the much-vaunted Oracle HCM Cloud is just Oracle Fusion tech stack underneath, that somebody else manages; in that case, I don't know if it's necessarily _fewer_ people that work on it, they just all work at Oracle datacentre instead of dispersed among the clients :) Haven't worked on Workday and the like yet to understand how "Cloudy" they trully are.
    - arethuza 8y ago
      Cloud based ERP/CRM do solve a lot of infrastructure issues - but they can also introduce a lot of issues (or opportunities depending on your perspective) - some integration tasks and data extract suddenly become a whole lot more difficult.
lunchables 8y ago
This is exceptional writing. It's concise, clear and engaging with a really clear call to action but not defeatist like so much writing regarding this issue.
- forrestbrazeal 8y ago
  Thank you, that's very kind!
finebalance 8y ago
I've just started my career in ML, and I already sort of feel that a majority of the work that people around me are doing can be done by an automated pipeline (like Python's AutoML) and throwing enough compute power at the problem. It's quite worrying.
- disgruntledphd2 8y ago
  Learn statistics and how to communicate them. That's going to be a long time being automated. For instance, if (for example) you model insurance data in the EU, you cannot use gender as a factor in pricing (even though it's effective). In general, the modelling/ML pipeline is the easy bit, the hard part is the data cleaning and figuring out how to translate a business problem into one that can be solved by data. tl;dr as many people have said about Comp Sci over the years, learn the fundamentals (statistics and experimental design) and you'll be in a much better position.
- ivalm 8y ago
  automl is pretty horrible if you're doing anything that is complicated, even at Google-scale compute power SotA DNNs are hand crafted, not auto-found. Feature engineering is also quite important for a lot of data types and that has too large of a phase space to probe with automl. At the same time if your problems are mostly solved by logistic regression/random forest/etc with simple features (ie you have well defined categories and/or states or your task is well solved in literature) then ML is not your value proposition, it's more data/business analytics (and autoML should enable you to provide your value faster).
cfontes 8y ago
I enjoyed reading it too, to the point and full of information. I really hate the current pattern of getting some interesting topic and nice data about and burring it into 20 pages of story telling journalism.
kyberias 8y ago
Ok, question: What exactly do you mean by this: '“senior application developers” who munge JSON in C#' Why did you choose to JSON and C# in particular here?
- forrestbrazeal 8y ago
  No particular reason, just trying to provide an example of low-value integration work that is likely to be subsumed by a service at some point.
- mx24 8y ago
  Because it's quite trendy to hate on all things Microsoft, especially in Silicon Valley startup culture. I'm a C# developer and have had other developers cringe in my face when I tell them I'm a C#/.NET developer.
  - mdmnk 8y ago
    I use VSCode and develop exclusively on Windows 10. I also have mac and linux machines. I think its better to be agnostic when it comes to language and OS.
toomuchtodo 8y ago
I don’t have twitter so I hope you don’t mind me engaging here. I would be very careful before dismissing AWS’ no code/low code project. Mulesoft, Microsoft Flow, and Zapier have a combined revenue of hundreds of millions of dollars a year in serving a market for establishing business logic workflow without code. I am only surprised it took AWS so long to move into the space considering the cross marketing opportunity to existing customers and their compliance capabilities.

NikolaNovak 8y ago

I've been in IT (mostly the "Legacy" kind - Oracle/PeopleSoft ERP) for about 20 years now, and I fully empathize with the article. I've subconsciously, then consciously, had to keep "moving up" on the value scale, partially due to outsourcing and partially due to automation. From sysadmin type to infra architect, but then made a curious curve that still catches me by surprise into tech lead and now find myself in basically management. I struggle to understand what will give me better skills and career for future: developing people and org knowledge and experience, as I do in my current role, which just feels generic and not particularly useful to somebody who used to gobble technical manuals for fun (but apparently market finds value in technically knowledgeable people who can manage and communicate); or go back to my love of technical and get back hands-on... understanding it's an uphill battle at my age and may be a losing one with the amount of automation and "AI being around the corner (forever;)".

arethuza 8y ago
If you have a decent understand of ERP systems and projects it might be easier and more productive to transfer those skills to more modern cloud-hosted ERP products than go back to more generic development skills?
- NikolaNovak 8y ago
  Thx; Agree, though I find many "cloud-hosted ERP products" are actually very traditional ERP products, just hosted by vendor (I remember the "Lightbulb" moment when I realized the much-vaunted Oracle HCM Cloud is just Oracle Fusion tech stack underneath, that somebody else manages; in that case, I don't know if it's necessarily _fewer_ people that work on it, they just all work at Oracle datacentre instead of dispersed among the clients :) Haven't worked on Workday and the like yet to understand how "Cloudy" they trully are.
  - arethuza 8y ago
    Cloud based ERP/CRM do solve a lot of infrastructure issues - but they can also introduce a lot of issues (or opportunities depending on your perspective) - some integration tasks and data extract suddenly become a whole lot more difficult.

lunchables 8y ago

This is exceptional writing. It's concise, clear and engaging with a really clear call to action but not defeatist like so much writing regarding this issue.

forrestbrazeal 8y ago
Thank you, that's very kind!

finebalance 8y ago

I've just started my career in ML, and I already sort of feel that a majority of the work that people around me are doing can be done by an automated pipeline (like Python's AutoML) and throwing enough compute power at the problem. It's quite worrying.

disgruntledphd2 8y ago
Learn statistics and how to communicate them. That's going to be a long time being automated. For instance, if (for example) you model insurance data in the EU, you cannot use gender as a factor in pricing (even though it's effective). In general, the modelling/ML pipeline is the easy bit, the hard part is the data cleaning and figuring out how to translate a business problem into one that can be solved by data. tl;dr as many people have said about Comp Sci over the years, learn the fundamentals (statistics and experimental design) and you'll be in a much better position.
ivalm 8y ago
automl is pretty horrible if you're doing anything that is complicated, even at Google-scale compute power SotA DNNs are hand crafted, not auto-found. Feature engineering is also quite important for a lot of data types and that has too large of a phase space to probe with automl. At the same time if your problems are mostly solved by logistic regression/random forest/etc with simple features (ie you have well defined categories and/or states or your task is well solved in literature) then ML is not your value proposition, it's more data/business analytics (and autoML should enable you to provide your value faster).

cfontes 8y ago

I enjoyed reading it too, to the point and full of information. I really hate the current pattern of getting some interesting topic and nice data about and burring it into 20 pages of story telling journalism.

kyberias 8y ago

Ok, question: What exactly do you mean by this: '“senior application developers” who munge JSON in C#' Why did you choose to JSON and C# in particular here?

forrestbrazeal 8y ago
No particular reason, just trying to provide an example of low-value integration work that is likely to be subsumed by a service at some point.
mx24 8y ago
Because it's quite trendy to hate on all things Microsoft, especially in Silicon Valley startup culture. I'm a C# developer and have had other developers cringe in my face when I tell them I'm a C#/.NET developer.
- mdmnk 8y ago
  I use VSCode and develop exclusively on Windows 10. I also have mac and linux machines. I think its better to be agnostic when it comes to language and OS.

toomuchtodo 8y ago

I don’t have twitter so I hope you don’t mind me engaging here. I would be very careful before dismissing AWS’ no code/low code project. Mulesoft, Microsoft Flow, and Zapier have a combined revenue of hundreds of millions of dollars a year in serving a market for establishing business logic workflow without code. I am only surprised it took AWS so long to move into the space considering the cross marketing opportunity to existing customers and their compliance capabilities.