Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Choose Boring Technology (2018)

boringtechnology.club

Post reply on HN
Print view View original

1–10 of 137 posts

Re: Choose Boring Technology (2018)

#3

Post by claytonjy »

Is boring technology even an option in high-velocity fields like ML/AI?

I can pick boring databases and boring servers but when it comes to model architectures, model weights, GPUs, serving frameworks, etc. there's been such massive improvements lately that it doesn't feel like anything has settled enough to be "boring".

If I need a language model, or an image classifier, or an ASR model, there are older options which might be "boring", but they're not nearly as powerful as the latest options. Does choosing boring ML tech mean sitting out on massive recent advancements?

Re: Choose Boring Technology (2018)

#4

Post by andrewstuart »

edit: Apparently "choose boring technology" is just clickbait, and in fact the guy advocates for all the nuanced technology selection advocated here.... so errr, apparently he in fact is saying "don't chose boring technology" or something, whatever.

I completely disagree with almost any blanket philosophy on technology.

Don't choose boring technology.

Choose technology that suits the context:

* choose what you like for personal projects

* choose tech your team has the skills with

* choose tech that won't be a problem to recruit people for (I'm looking at you, Ruby On Rails)

* choose tech that gets the job done

* choose tech that the industry is moving towards

* choose tech that levels up your team's skills and makes them feel excited that they are learning and advancing

* choose tech that allows you to meet the priorities of the project (your project does, define priorities, right. right?) security/availability/maintainability/performance/compatibility/whatever

* choose tech that ensures your team and company won't be left behind when market skills move on

* choose the cutting edge if you think that's the right choice for some reason

It seems pointless to me to say "choose boring technology".

Re: Choose Boring Technology (2018)

#5

Post by dgrin91 »

claytonjy wrote: post #3

Is boring technology even an option in high-velocity fields like ML/AI? I can pick boring databases and boring servers but when it comes to model architectures, model weights, GPUs, serving frameworks, etc. there's been such massive improvements lately that it doesn't feel like anything has settled enough to be "boring". If I need a language model, or an image classifier, or an ASR model, there are older options whic…

99.999...% of us don't work or seriously touch in AI in any way. At best you make an API call to something that might do some AI, but at that point its just another service.

So for the vast majority of people doing dev work, boring choices are correct. I've had a lot of junior devs come to me with proposals for microservice architectures that can autoscale, be orchestrated, have 0 downtime with red/blue deployments, etc, etc, when at the end of the day it could be done with a boring, single server (plus second for resiliency, but not needed for scaling) with just a few endpoints. It will be done much faster & simpler, which means it will be cheaper and less likely to fail.

My guess is that the same is true in a lot of AI work. Yes there is core stuff that needs to be performant, scalable, and modern, but I'm sure there is also a ton of boring old regular internal corporate infra that can be done 'boring'.

Re: Choose Boring Technology (2018)

#6

Post by ryanschneider »

claytonjy wrote: post #3

Is boring technology even an option in high-velocity fields like ML/AI? I can pick boring databases and boring servers but when it comes to model architectures, model weights, GPUs, serving frameworks, etc. there's been such massive improvements lately that it doesn't feel like anything has settled enough to be "boring". If I need a language model, or an image classifier, or an ASR model, there are older options whic…

I think the whole point is that you only have a limited “cutting edge budget” so if you are somewhat forced to spend that budget on cutting edge ML models then don’t overload your team _also_ using cutting edge languages or infrastructure or databases.

So save your budget for what will really make your product unique instead of “wasting” it on things that in the end won’t give you a real competitive advantage.

Re: Choose Boring Technology (2018)

#8

Post by Pannoniae »

While I see how this makes sense, I don't like this. If most people follow this strategy, we will end up with utter stagnation, or even devolution.

"Boring" technology also has many hidden costs - most importantly, opportunity costs of not having several times higher productivity due to the warts and quirks of the boring tools. Sure, with enough practice one can work around them but every backwards issue takes a good chunk of productivity out of the project.

I get the argument of "if it's not mainstream, where will we find programmers" - but that's more of a cultural issue of wanting expert programmers out of the box instead of investing in them and training them.

Ultimately, this line of thinking is like saying 640K RAM is enough for everyone. Sure, you might say C++ and Go is enough for everyone, but there are many unrealised gains left on the table.

Re: Choose Boring Technology (2018)

#9

Post by tedunangst »

andrewstuart wrote: post #4

edit: Apparently "choose boring technology" is just clickbait, and in fact the guy advocates for all the nuanced technology selection advocated here.... so errr, apparently he in fact is saying "don't chose boring technology" or something, whatever. I completely disagree with almost any blanket philosophy on technology. Don't choose boring technology. Choose technology that suits the context: * choose what you like f…

Open the link again and scroll down.

Re: Choose Boring Technology (2018)

#10

Post by jvans »

claytonjy wrote: post #3

Is boring technology even an option in high-velocity fields like ML/AI? I can pick boring databases and boring servers but when it comes to model architectures, model weights, GPUs, serving frameworks, etc. there's been such massive improvements lately that it doesn't feel like anything has settled enough to be "boring". If I need a language model, or an image classifier, or an ASR model, there are older options whic…

Yes. Use GBTs and off the shelf language models you don't train yourself.
Post reply on HN