[ad404b8a372f2b9]

I like the concept but the curation would have to be done ruthlessly by experts, otherwise you'll get the same SEO optimized garbage we find on google. To see Machine Learning Mastery at the top doesn't inspire confidence in your curation process. It's a blog that's emblematic to me of low quality content pumped out for SEO purposes.

[mlent]

thanks for your feedback (OP here). yeah it's tough, i try to give anyone a chance to rank who has more than just a few posts.

what I didn't expect about that site is that it gets shared A LOT on twitter. still trying to refine the algorithm :)

thanks a ton for checking it out and sharing your feedback!

[ad404b8a372f2b9]

It occurs to me you could switch the curation and trend gathering steps to minimize the amount of work on your end. Instead of going "twitter trend -> filter out the bad ones with in-house experts" you could go "follow experts -> find out what trends among them". That's my usual workflow when using twitter.

[mlent]

that is already what i'm doing pretty much. i accept blog submissions and try to look for signs of quality, and then monitor for mentions of their sites.

as you suggested though, i didn't know the ML space enough to know that Machine Learning Mastery publishes daily and has an army of people who like and retweet ANYTHING they publish.

so it makes me wonder if i need to have some kind of dampening effect or how i can adapt the algorithm to handle that.

[zerkten]

Get some experts in the area to rank your existing content and suggest weightings for their recommended sources. This is the less from launch after launch of really great sites that aggregate information.

Let's take Stack Overflow as an example. Jeff found a small group of experts and expanded it. They seeded both questions and answers. They didn't bring on just one or two experts though, they brought on enough to ensure a good distribution (not perfect) of viewpoints and then reviewed before expansion. They kept repeating this and didn't optimize for just one kind of developer (Django over Java.) All segments of developers tended to need the same features, but it would show up with one segment first. Getting answers on some topics wasn't possible until the product was more mature. Kill crap ruthlessly like SO did with downvoting and moderator-led deletion.

If you are building an ML model then you are going to need to find a range of experts and either seed from what they are sharing, or create a review system. You can reward people with kudos on a contribution page, donations to open source projects or charities (even on behalf of a group of them), or find another way to motivate them. It just needs some hustle, but you've got to forget about purity, be open about how your model works, and iterate.

[mumblemumble]

Perhaps there could be a ranking that takes posting frequency into account, and lets people decide whether they want to see higher- or lower-volume blogs.

Perhaps not a straight ascending or descending by frequency. For me, the sweet spot seems to be no more than once a month. More if it's one that does something like a weekly post aggregating interesting articles from other blogs.

[mlent]

That's a cool idea! Adding an "advanced search" is something I've been thinking about.

Gives people more control over their personal "algorithm" and what they value most when reading personal blogs.

Hadn't thought about frequency but that could definitely be something you could automate based on RSS. Thanks for the idea!

[pc86]

Why would you penalize a blog with 3 authors who publish 3x the articles of a personal blog? Why would you penalize higher volume at all? Trying to ascribe anything close to quality based on volume is like trying to grade developer productivity and/or skill by LOC.

[grep_name]

He's not asking them to be penalized, just to let people choose what their speed is. I get the same problem on mastodon: there are users there whose posts I really appreciate and find interesting but I had to unsubscribe from because their posting frequency flooded out everything else. I just check up on their feed separately every once in awhile instead

[mumblemumble]

I'm not asking for some strict formula, more some ability for people to describe what kinds of blogs they're after in more detail in a way that's flexible, but not overly fiddly to use.

I'd possibly put a shared blog under a similar category as the ones with weekly roundups. Which is exactly why I threw out that example - I don't want hard-and-fast Google-style rules, because I would expect that to work out about as well as hard-and-fast Google-style rules ever does.

[mycall]

Is there plans for I18N for the content buckets?

[pc86]

I've purchased whatever Jason calls the big bundle from ML Mastery and the books are pretty much the same. I get the impression the primary goal was "produce a lot of pages," while "impart information" was secondary.

That being said, it looks like points are awarded two ways:

* Counts the least: Retweets and likes of a tweet containing a link to an article on the blog

* Counts the most: Replies and mentions of an article on the blog by other people

So I'm sure it's possible to game this, and it seems to bias toward incumbents (but what doesn't), but at least it's trying to be a little better than "here's a list of blogs I like."

[sixhobbits]

Machine Learning Mastery has working code samples. That already puts it in the top 1% from my perspective and I definitely relied heavily on Jason's blog posts while doing my NLP masters.

He definitely has a far more commercial focus than many other bloggers but this seems a bit too critical to me.

[pc86]

I mean I spent hundreds of my own money on the bundle so it's not like I think he's a hack or fraud or anything of the sort. But there's a decent amount of duplicated content.

[kevas]

Statistics on the complexity of replies?

[stiray]

Exactly. And SEO optimized garbage is already what we get currently.