dcsan

13 ms·

91.
by dcsan 6y ago
it's a different role. he signed on to be a media exec and cut licensing deals. suddenly he's in the middle of a trade war and politics with some of the worst people on the planet. "fine people on both sides"
92.
by dcsan 6y ago
his usefulness expired, and now HQ want to control the nego directly. Just like any foreigner used as a local hired gun.
93.
by dcsan 6y ago
2nd for FOAM. Its really awesome. this VSCode as a platform is taking off. There are other markdown renderers like dotgraph that can easily be embedded in your visuals using other extensions. And then all the power of VScode editing on top.
94.
by dcsan 6y ago
agreed. it's all the lockin of a non-IMAP service, but with a couple of UX things that seem nice (?) and others can easily add. Maybe google won't rip off quite as blatantly as facebook and snap but... something similar will show
95.
by dcsan 6y ago
I wonder if a "dispatcher" model might be interesting for this, like an old style minicab/police service. Then people could use squawk-scanners for the channels they want to listen into. It could get annoying if its on all da
96.
by dcsan 6y ago
looks very polished. but depends on an XML fileformat. While basically text the appeal of something based on simple to edit .md files is strong...
97.
by dcsan 6y ago
It depends on the editor, but esp for markdown pure text, WYSIWYG is a thorny topic - see slack's recent problems with their IME. Typora has a nice feel and switches from editor to preview as you move to the next area in a document. Ob
98.
by dcsan 6y ago
This looks really cool, but: Is there a reason this doesn't work with plain markdown links? I have a local working kbase already as a bunch of markdown files. If I use the obsidian markdown importer I guess it will destroy any links in
99.
by dcsan 7y ago
This does use spaCy for english tokenization, but has a bit wider support for languages in general. eg spaCy has just finished their Chinese models in beta Discussion: https://github.com/howl-anderson/Chinese_models_for
100.
by dcsan 7y ago
you can absolutely download the models locally: # download best-matching version of specific model for your spaCy installation python -m spacy download en_core_web_sm
101.
by dcsan 7y ago
Can someone ELI5 where all the costs are to flip from black to red like this? Are they not including things like salaries to calculate gross margin? It seems those are the COGS of an IP based business, not raw materials ---- GAAP gross prof
102.
by dcsan 7y ago
lots of comments there! / xml from char
103.
by dcsan 7y ago
It's maybe hard to know "what you love" but it's easy to know what keeps floating back up in the top of your brain, even if rational you says it might not be the best use of your time. Then again if "Dance Monkey&qu
104.
by dcsan 7y ago
“I spent half my money on gambling, alcohol, and wild women. The other half I wasted.” - WC Fields
105.
by dcsan 7y ago
I've started writing down those ideas of interest so I can see if there are themes, or better still a meta-idea that connects some cluster of them, to start work on.
106.
by dcsan 7y ago
even though many projects actively try to dissuade "user" questions and direct you to SO. Now they'll maybe send you to the relevant discussion. This is closer to the "source" for answers specifically linked to one
107.
by dcsan 7y ago
Zeit's was the most active spectrum community I was part of, interesting they're the lead tester community for this new GH feature. Spectrum fell between a number of posts, it was never really "chatlike" so it looks the
108.
by dcsan 7y ago
so you didn't have a word near penultimate and then pick some letters from it? It's really just open-ended what the users should choose? That does make it a lot harder...
109.
by dcsan 7y ago
btw what did you like most about using the magnitude format? lazy loading seems good, depending on if it takes time to load the sqlite DB at each startup. That's one of the major pains of working on anything with word embeddings is tha
110.
by dcsan 7y ago
these are neat, it would be fun to put a reactJS frontend on some of them!
111.
by dcsan 7y ago
Yes I would look to fasttext first, they have a huge library of vectors here: https://fasttext.cc/docs/en/crawl-vectors.html Actually facebook research are pretty good at gathering things, their ParlAI project als
112.
by dcsan 7y ago
You could have fake interviews with clients as chat bots. It wouldn’t have to be a great bot as clients often ignore your question, give very vague responses. As long as you had a few responses lined up that had to be triggered by asking th
113.
by dcsan 7y ago
Chinese (and japanese) use 4 digit separated factors, so it's easy to make these kind of mistakes. eg 100 ten-thousands not 1000 x 1000s. Even after living there awhile, converting big currency numbers is a pain.
114.
by dcsan 7y ago
this looks really interesting as a format; I agree that Twine is hard to engage with, the editor code is a mess and the generated format isn't clean. But it is quite fun and quick to use as a visual story "map" editor. For yo
115.
by dcsan 7y ago
is there a mirror somewhere? article seems interesting.
116.
by dcsan 7y ago
sidekiq! Interesting wenbin also wrote a task queue system... must be something about that type of engineer that leads to this pragmatic tech strategy.
117.
by dcsan 7y ago
It seems wenbin is also the lead developer on ndkale, https://github.com/Nextdoor/ndkale/graphs/contributors so he would certainly be capable of using a more esoteric stack. I wonder when developers are work
118.
by dcsan 7y ago
not quite related but ... any recommendation on a good open source mud? most of the projects I've tried have been very raw, or unmaintained. eg https://ranviermud.com/
119.
by dcsan 7y ago
finance is the meta-game. "All problems in computer science can be solved by another level of indirection"
120.
by dcsan 7y ago
GP grand parent https://news.ycombinator.com/item?id=8363625

More ›