How to Build Your Career in AI [pdf]

I think it makes for a great ironic read if you replace AI with Blockchain and imagine this was released 5-10 years ago. Not to snide but just find it funny comparison.

Re: How to Build Your Career in AI [pdf]

#5

Post by andyjohnson0 » Mon, Oct 28, 2024, 11:01 AM UTC

This is from 2022 [1]

As someone who's tried and mostly failed to get anywhere in the AI space these past few years, I'm wondering how current AI practitioners perceive this advice two years on?

For prospective new products it feels like effective moats have been constructed - and you mostly have to pay to use someone's api. And on the career perspective its hard to get experience now without pre-existing experience.

Wondering how others see this.

[1] https://www.linkedin.com/posts/andrewyng_how-to-build-your-c...

Re: How to Build Your Career in AI [pdf]

#6

Post by agomez314 » Mon, Oct 28, 2024, 11:06 AM UTC

I think, for most people, knowing the right person or arriving at the right workplace at the right time is more important than most of the advice outlined in this doc. How many talented people have been put down by a bad boss? Or extremely hardworking, smart engineers that cannot find a job in today's market? Yes, learn fundamentals (learn them deeply!), do projects, but don't expect to have a successful career just because you did these. One may follow the other, but it is never guaranteed - focus instead on trying to live a good life and be kind to others.

Re: How to Build Your Career in AI [pdf]

#7

Post by JTyQZSnP3cQGa8B » Mon, Oct 28, 2024, 12:14 PM UTC

romanovcode wrote: post #4

I think it makes for a great ironic read if you replace AI with Blockchain and imagine this was released 5-10 years ago. Not to snide but just find it funny comparison.

Blockchain or any other skill. It’s perfectly generic, and the stock photos add a professional flavor to this useless guide.

Re: How to Build Your Career in AI [pdf]

#8

Post by davedx » Mon, Oct 28, 2024, 1:41 PM UTC

andyjohnson0 wrote: post #5

This is from 2022 [1] As someone who's tried and mostly failed to get anywhere in the AI space these past few years, I'm wondering how current AI practitioners perceive this advice two years on? For prospective new products it feels like effective moats have been constructed - and you mostly have to pay to use someone's api. And on the career perspective its hard to get experience now without pre-existing experience.…

From a "learning the basics" perspective I think it's fine? It could maybe have more emphasis on transformers I suppose.

I think to get somewhere in the AI space you usually need to work somewhere that's solving real problems where AI can help. I know that sounds trite and reductive but it's also where the success lies IME.

Re: How to Build Your Career in AI [pdf]

#9

Post by andyjohnson0 » Mon, Oct 28, 2024, 4:56 PM UTC

davedx wrote: post #8

andyjohnson0 wrote: post #5

This is from 2022 [1] As someone who's tried and mostly failed to get anywhere in the AI space these past few years, I'm wondering how current AI practitioners perceive this advice two years on? For prospective new products it feels like effective moats have been constructed - and you mostly have to pay to use someone's api. And on the career perspective its hard to get experience now without pre-existing experience.…

From a "learning the basics" perspective I think it's fine? It could maybe have more emphasis on transformers I suppose. I think to get somewhere in the AI space you usually need to work somewhere that's solving real problems where AI can help. I know that sounds trite and reductive but it's also where the success lies IME.

> I think to get somewhere in the AI space you usually need to work somewhere that's solving real problems where AI can help. I know that sounds trite and reductive but it's also where the success lies IME.

I think thats very true, and not trite or reductive at all.

I used to work for a data-heavy SME and tried to get the senior leadership interested in exploring AI in some commercially useful problem spaces. The main issues I had were not the tech, but resourcing and getting sufficient buy-in to work on it for more than a few sprints. In the end I failed on both counts.

In career terms its like "you can't get there from here". I wish there was a howto for that.