Totally agree! People's willingness to tell you about their thoughts is the N.1 way to get a call started.
Now imagine that group of students opened the conversation with: "I saw you raise your hand in C101 this morning, and thought your question was very interesting. What made you think of that?"
I would think the same if I didn't experience the opposite. To be honest, I would start with an email / linkedin mostly, but even cold cold calls work. I remember doing exclusively cold cold calls for one side project (airbucks.io) - and I got an awkward early ending to the call maybe 1 out of 5 times. Which doesn't faze you much when you've just had a blast for 30 minutes with a potential customer.
Any type of software development services and business where the end result is some form of data (and you can send it as a spreadsheet) - i think this covers almost anything. On short thought - the only big exception I could find are 2-sided marketplaces.
I totally agree - we all know of the extremely irritating cold calls. However, in my own experience, if you research the person you are calling well, and follow a few rules of thumb (be genuine, aim for learning) the feedback from people is surprisingly positive! (I'd say maybe 1 out of 10 is a slightly irritated response)
Here (in Germany) it's legal to call businesses.
And also: It doesn't have to be calls - it could also be email or linkedin - and yes! also here it works if genuine.
"Cold call ~20 people who might be good customers."
IMHO this is probably the first! thing you should do. To my own surprise people will give you money for your service even if you don't have a websites (and name, logo, slogan or anything else), and you send your 'product' via email.
And 3 out of 4 ideas don't survive these 20 calls - so you'll save a lot of time if you sell first and build later.
Emotionally that's not easy - of course - but it's what you will be doing all day anyway if it goes well, so why not start early?
Behind AI I would always expect at least Deep Learning. Machine Learning I use for everything that learns it's own decision boundaries. When it's humans teaching a machine, I'd call it simply automation or expert-system.
Although a lot of "teaching" still goes into feature engineering ...
Hi, Markus, Founder of DataRevenue here - we can't open the original sheet for editing as that would put the employees data at risk. So here's another sheet for companies to post and edit directly - https://docs.google.com/spreadsheets/d/1JDx5acZPvdmmSeMyEr77...
I also think thats justified. A newer, and more specific term I particularly like is "Machine Learning Engineer", which will probably soon be recoined to "AI Engineer".
We (www.datarevenue.com) basically have to use "AI" now to make it clear what we do. Something that would have made me feel awkward just 3 months ago.
Independently - I was passed the list as an effort by SoundCloud to increase exposure. No one til this point had thought about posting it on HN. No affiliation with SC and not in the recruiting business.
Now imagine that group of students opened the conversation with: "I saw you raise your hand in C101 this morning, and thought your question was very interesting. What made you think of that?"