My approach is to provide 3 topics of conversation (no right answers just talking): Let's say the person is applying to be a dev in product P. 1 how would you build an mvp for P? Which stack, what are the main issues, how would you deploy etc (usually I ask if it were a side project , so I can see which technologies are in the intersection of fun and reliable for this person)

2 what would be a moonshot feature for P? (If you had a perfect AI lib or unlimited processing power or devices spread all over the world)

3 if you had to find someone to take over your job (for you to manage the new one), what would you look for?

Rust maybe, python definitely no and Julia I don't know a lot.

How is Python definitely not "safe, productive and easy to use". It has faults, but these three are not them.

is there any good terminal assistant? something that accepts these kinds of queries but with in a cli, preferably without the need of constant internet connection

Mycroft has a curses CLI: `start-mycroft.sh cli` or `mycroft-start cli` if services are running, `start-mycroft.sh debug` or `mycroft-start debug` to start services and go straight into the CLI.

I learned french in 3 months aprox. while living in France. Ive arrived there without much knowledge of the language. First weeks I was feeling like an alien. Then I started to understand but couldn't speak. Then I started to use translate and dictionary to build some phrases beforehand to be able to be polite and ask the questions I wanted. After 3 months something clicked during a talk I was watching, I not only began to understand without having to think but was able to ask questions to the French guy on my side about the talk (it was about a ML algorithm). I am a native speaker of a Latin language (Portuguese).

I strongly disagree with the author. I've introduced tens (maybe hundreds) of people to code and starting with code (instead of a GUI) and preferably lower level (like C or VB) code is the best way to keep people motivated. What I've found is that people tend to settle in the less difficult path. After teaching python, no one cared to learn C, after teaching things like dreamweaver, no one cared about html. Even when they realised the benefits, they already had a tool that was good enough. If your objective is to simply make people more maker, ok, but I yet have to meet someone that would follow the described path.

Author here. Thanks for sharing your experience. But how do you know whether Python or C served his needs better? For all I know it could be Python -- you can build a web server using Django, write scripts, do machine learning, do computer vision using OpenCV... These might very well be more useful to that person than what C lets you do. Nobody should learn C because you think it's important; they should learn it if, and only if, it's the best tool for what they want to do.

possibly migraines? it started when i was around 23 and I had the same symptom (unreasonable back/neck pain). I just started to have headaches after visiting a neurologist and getting medication (pamelor), in the first month of medicine, the back pain stopped. after some months of headaches (and medication) most of it stopped. try to see if the pain correlates with coffee, sleep and fat food. everyone has certain triggers and they may take a while to appear. if I'd drink too much coffee, I'd have migraines in the next day. eating fat food on friday would just appear as pain on sunday or monday...

In my company, side projects are divided into two groups: products or toys. Both can be done during work hours without problem (if you continue to deliver your sprints, of course). Your work should always be deletable (you should keep version control somewhere non local) and your peers should always be able to see your code (asking or not, If it is in a company computer, everyone in the company has access to it). We also incentivize presentation on new technology or topics that emerged from the projects and usually people pitch each other their projects for new contributors. Projects considered products (that the programmer has started expecting profits at some point) should consider our non compete policy: if the product is in the same field as one of our products, our company has the right to be the first investor, with a minimum of 5% and cap of 15%. The valuation process is a bit dense to explain here, but several ex coworker are now partners with their own companies. The system works well but we have to watch out for talent retention. Usually the new companies are SaaS that replace part of our needed logic and cost us less to use than our ex coworker salary.

Thats a very cool model. And a really good point here:

> Usually the new companies are SaaS that replace part of our needed logic and cost us less to use than our ex coworker salary.

Its a win-win to have the company as the new founders first customer too.

I feel like this is the way to do it. It's hard enough to start a business as is, so not that many people are going to actually do it. People that want to do it are going to do it anyway eventually.

This is a very good model.

IMO it's github fault to use this star gimmick. Likes, stars, upvotes... all of these are dopamine hits disguised as curation or rating... The UX designers or engagement/growth hackers or whatever that plans and are responsible for these should try to predict the consequences of a system so easily exploitable to be implemented. Most successful MMOs have whole economies balanced, how can a 1 variable (can be more considering forks, contributors etc) economy be this open? Of course someone would exploit it. Github could hide forks and contributors but it could eliminate stars, since it brings nothing to the table more than virtual ego. Stop the likefication of software. We need to go back to free as in freedom, not free as in gratis.

I agree completely. Github is primarily a social network site. Code on Github is what tweets are on Twitter, or images on Instagram. Grist for the social media attention mill! And just like every other social media site, it's being gamed by the egotists and the self-promoters and the marketers and the fakers.

First, I also had trouble with XCode. About the regex in vi: there is a technique of writing/maintaining code based on substitutions instead of inserting/deleting. It can be very powerful but you have to "write" your code thinking about the patterns and maintaining some kind of structure. I am not saying that it is good or recommended. It's interesting. I coded for like 5 years like that. I was maintaining a codebase completely solo. When working with a not very synchronized team or having tools like resharp or omnisharp etc avaliable, I don't think it would pay off.

You can surely think about a model of sandboxing with public and private (even internal for company wide) restrictions.