22 comments

[ 3.6 ms ] story [ 61.3 ms ] thread
Why isn't the description the name of the product?
Why isn't your biography your name? Because a name is way more valuable when referring to specific things. It doesn't draw out conversations, lets you distinguish between multiple similar items, and serves great in making sentences easier to parse by making its parts more distinct. I'd much rather have a conversation that goes like

> Have we considered Cloud Vision yet? I think Cloud Vision has the best image recognition results out there.

instead of

> Have we considered Image Recognition and Classification? I think Image Recognition and Classification has the best image recognition results out there.

"Why isn't your biography your name? " because the name come before the biography.

In a products case you hope that google has a vision about what the product is before naming it..

The list is very useful..thanks..

A name is a primary key tuple of a subset of the metadata that describes the named object. How much metadata should a name contain?
Is anyone finding medium.com to be unusable when not logged in? From a big banner at the tops and bottom and a huge popup on every visit, I think it's time for people to move content elsewhere.
Same here, they're quite aggressive now about how many articles you've read. Not cool.
Slightly off topic, please forgive me.

What kind of cloud product should I get if I want something that's equivalent to a Raspberry Pi on my own home connection, to run small scripts and maybe some small server applications without any real traffic? The load could be considered insignificant, but I'd like it to be available 24/7, and I want to install my own programs and libraries.

Is it as simple as a VM? What's a good place to get one, Digital Ocean? Being able to use my own domain name for access would be a plus.

I have a DO box... the cheapest one. It works great. I have a personal website on it. For what you want, I don't know how you would have any issues. I have never really tried anything else ( I use amazon EC2 free tier which is cool if you want to deploy different environments but DO works just fine.)
Yea, you're looking for a simple VM. A cheap Digital Ocean droplet would be perfect. They are cheap, reliable, and incredibly simple to launch. DO has a nice few offerings you can layer in over time if needed but getting that single VM up and running is still incredibly easy and fast.

I do highly recommend checking them out.

DigitalOcean has been mentioned; there are providers of rather cheap VMs. For example hetzner.com and ovh.com are rather popular in Europe, and then there's linode.

If you are fine with ARM hardware instead of the usual amd64, there are rather cheap and high-powered offerings in this space, for example https://www.scaleway.com/pricing/ (typically, the small VMs start around 1GB RAM and 1 core, whereas here you get 2GB RAM and 4 cores at a comparable price).

Vultr offers $2,5 servers on few locations (I think NJ and 1-2 more). Those are amazing for testing and small work. They are comparable to their $5 servers from a year ago performance-wise. Great value.
Google Cloud gives you a f1-micro VM for free without a time limit. This about equal to a RPi.
That's not available in Europe for some reason, at least not for personal projects without potential economic benefit.

> In the European Union and Russia, Google Cloud Platform services can be used only for business purposes. If the sole purpose for which you want to use Google Cloud Platform services has no potential economic benefit you should discontinue your use of the service.

https://support.google.com/cloud/answer/6090602

You should have a look into serverless:

https://serverless.com/

I use the nano instances on AWS as well as they are very cheap to keep running. If you are new to AWS you can also run a micro for free for a year.

I’m surprised no one else mentioned Scaleway given your description of something like a Pi
Um it sorta didn’t work by the second example Cloud Job Discovery -- ML Job Search and Discovery That’s either five or six words depending on whether you count machine learning or machine-learning as one or two words
Don't count the "and" and you hit the target of 4 (if you accept that ML is one word).
Am I the only one who finds many of the descriptions less descriptive than the title of the service? I've always found GCP pretty on-point with their titles