Very unfortunate, MPC-HC was so simple to use, has a slick UI inspired by MS Media Player and to quickly review/seek videos it was the very best. (much faster than VLC for that task)
I have only been exposed to Dynamics CRM (Dynamics CRM 2011 through to Dynamics 365) I must admit. I can't imagine the experience for these other product strands to be much better for developers let alone users though.
Yes, we're analyzing the titles. They're vectorized and fed into a LSTM net which was trained on a manually tagged training set.
The set is not very big yet but yields good enough results for the initial proof of concept.
The samples were selected randomly at the beginning. Tags were thought up on the go, trying to generalize into broader categories.
After the initial tagging, we counted the number of samples for each category and grouped similar underrepresented tags together + added additional samples for tagging that could go into the smaller categories just by filtering those who matched specific keywords and further hand-picking them.
We initially tried training the classifier only with GitHub based samples and using the user-given tags from there. Although we grouped the tag base into a reasonable number of distinct categories, the way how GitHub users tag their projects turned out to be just too inconsistent and often unrelated to the titles, so manual tagging was seen as a better option for getting decent results fast enough.
If you have any more specific questions feel free to drop me a mail to arturs@finch.io
For those of us in industries at the fringes (but still of interest) to HN, want to have an 'Other' category? :-) We make scientific/genetic software - and I'm not sure where we might have fit in your sidebar.
Getting to the front page via 'Show HN' was very helpful to us. It'd be nice (for others too) to be able to both replicate that success, and soften the blow when you get a grand total of 2 upvotes.
> Erlang code is single threaded. Each Erlang "process" is an isolated, share-nothing single sequence of instructions, and you don't use semaphores, locks, or critical sections in writing Erlang code. The BEAM VM is implemented as a multithreaded process, but this is not exposed to the Erlang code
The same is true for PHP.
No wonder Facebook developed HHVM for PHP and also uses Erlang with WhatsApp and other parts.
"Complex problems don't have simple solutions. It could be a series of sticks: a vacancy tax, an increase in tenants' rights through guaranteed leases and binding arbitration over rent increases (as outlined in the Small Business Survival Act). Or it could be carrots: tax breaks to small businesses to help ease the pain of high rents, or streamlined registration and a reduction of regulatory hoops for certain types of retail tenant. [...]"
And some possible solutions to fix this complex problem
I'd like the first two to apply to apartments, any rent-able property really.
After that is the development stick. Tax property at a minimum based on the surrounding average property tax rates. This will encourage proportional investment in areas and prevent hold-outs who are waiting for some hypothetical from sitting on a community asset forever.
What if for some reason the property is really worth less than the surrounding properties? Seems like a recipe to have properties that barely pay for themselves, which discourages investing in renovations, etc.
You don't really need "AI" for this. For this "triage" use case a good expert system/decision tree is more than enough. Think of it like akinator for a given domain.
Hand-building the tree to be efficient is actually the hardest part.
Edit: yes, this is probably "AI" to some people, but AI only exists as long as we don't understand it. This kind of tech has been well-understood since the 70s.
There's a fair bit of AI, NLP, etc. The system starts out with decision trees built by/evaluated by medical professionals and learns what conversational elements are required to reach a diagnosis. This isn't just a trivial set of hand-built trees.
I am with you. In many use cases HTTP 1 is fine. And as we have seen with the Hearthbleed bug, if you cannot devote resources to keep your HTTPS up-to-date all the time 24x7, it makes your server and the user sessions more insecure. And often providing HTTP and HTTPS makes sense too. So I am against labeling HTTP as "evil" or legacy. We should look who has an interest in doing just that, and why they are forming initiatives. Beside all that pretty much all ad-networks serve HTTP ads, so if you want to loose much of your ad revenue, please go with the HTTPS-only route and face the real world surprise.
Is there a way to op-out of this on iOS and macOS? That's really scary.
ML and AI on OS-level should run decentral on the device itself, and don't leak data at all. The spirit of the 1990s was that way, and we older desktop software works fine that way (even on Pentium 1 hardware), so it would run like a piece of cake on a modern smartphone.
The "differential privacy" technology may sound good, but without an independent audit who knows how good it works.
