500 days on Duolingo and it has felt like it was produced by ChatGPT 3.5 this entire time.
500 days of learning various bits & pieces and not being able to have a simple conversation - but I could probably say "There is a monkey in his backpack" if pressed hard!
I used to hate learning from actual textbooks as the conversations felt "dumb" or "forced" but that dumbing down as at least justified by having to progress from zero. Duolingo doesn't feel "plain dumb" but "weird dumb".
So yeah, if they replace their contractors (who must've used the cheapest models) with O3 or o4-mini-high or whatever, it should be an improvement!
Seasoned Product Engineer with 20+ YOE. Looking to help NYC founders scale beyond the MVP as your interim greybeard. I will clean up your code base and set you up for growth. Can help with infra (cloud and on-prem), performance, scalability, and agility. Past clients include Bloomberg, AMEX, Volkswagen. Tech stack includes Node.js, .NET, PHP, SQL and NoSQL.
Helping with maintaining and modernizing legacy code bases - PHP, JavaScript, .NET. As a seasoned generalist, I'll make sense of the weirdest stack you might have, document it, modernize it, and extend its ROI well past 2024.
Areas of expertise include: web performance, API integration, app deployment (on-prem and in the cloud), line-of-business desktop apps (.NET/WPF)
A good way to get started is not to ask "How do I become a freelancer/consultant" but rather "what valuable business outcomes can I deliver with my technical skills" and then find clients who have those problems and solve them.
Yes, this ignores the hard part of "finding those clients".
I know people with expertise in improving website performance, for example, and the mindset shift here is not to think about response times but rather improvement in conversion rates, decrease in shopping cart abandonment rate, etc. Clients don't really care about caching or asset preloading but they really, really care about squeezing the extra dollar from their website visitors. Help them make that extra dollar and you get to keep some portion of it.
Once you establish this foundation, getting to the right clients is easier than if you put yourself out there advertising "will write Python code for $$$".
So for instance, I wouldn't necessarily attend technical conferences / meetups but perhaps conferences where, I don't know, ecommerce website operators discuss issues that are critical to them. Those are your potential clients who can be perceptive to the right pitch. I think that technical conferences are a better fit if you're looking for a job and selling your resume.
Having said all of this, when I started out as a freelance dev, I did use Upwork to get some experience and reviews that I could later repurpose as testimonials on my website.
There's nothing wrong with hands-on freelance coding to get started somehow. It can be hard to define what specific business outcomes I could deliver with technical skills/existing resume, so getting some real-world experience can help with refining that. Always be on a lookout for that answer: once you understand "why" your clients keep paying you (hint: not for the code itself), you can find your niche and improve your sales pitch.
Finally: yes, it's critical to get out there and connect with people! Be it on LinkedIn, email, forums, Discords: you won't get clients unless you talk to people. Talking to past connections could also be a good way to find contract opportunities esp. if you've kept in touch over the years. Being a solo dev/consultant is 80% people work and only 20% technical work.
PS There's another way: you can ignore all of the above and apply to contract jobs (1099) using recruiters to get to clients. That's very similar to employment (clients tend to treat you just like staff) but you can potentially get better rates. I've been doing this for years with "success" but frankly, it's only marginally better than employment plus you get no benefits like vacation days etc. Real, actual consulting/freelance work can be a lot more satisfying if you can make it work!
https://github.com/teekay/dictum
Currently dogfooding and evaluating whether it helps in the long term or not.