dado3212

10 ms·

91.
by dado3212 4y ago
@dang can we pin this?
92.
by dado3212 4y ago
Previous discussion: https://news.ycombinator.com/item?id=34565909
93.
by dado3212 4y ago
Would you put holdouts of good features in the same bucket as negative testing? Let’s say someone develops an ML model that cuts down spam by 90%. It’s launched to all but 1% of users, so that you can consistently measure that the spam redu
94.
by dado3212 4y ago
Is there a way to see what the optimal path was after you’ve solved it? Or just the optimal number?
95.
by dado3212 4y ago
Are these photos in true color? What do the RAWs look like? I can't tell if the coloring is real, or if it's more along the lines of astrophotography where the colors are fake, but illuminate something key about the photo (in this
96.
by dado3212 4y ago
Is there any aggregation of near misses? How often do things like this happen?
97.
by dado3212 4y ago
Would also be awesome if there was deterministically only a single solution, though that may be hard to create.
98.
by dado3212 4y ago
Is there an easy way to look all of them up?
99.
by dado3212 4y ago
This feels like ChatGPT.
100.
by dado3212 4y ago
The link I think you’re missing is the strong impact podcaster revenue has on what podcasts are produced and the audience that consumes them.
101.
by dado3212 4y ago
I don’t have anything against RSS, I think it’s a useful protocol that fills a niche. I do however think that it’s just that: a niche. If you read through the other comments on this and other historical posts it’s presented as a panacea t
102.
by dado3212 4y ago
RSS has got to be the thing that Hacker News loves the most that every non-tech person on Earth doesn't care about. So many techies love this model of designing their feed and collating what content they consume, while the quickest gro
103.
by dado3212 4y ago
The podcasters care? Because if the advertisers can't quantify ROI and audience, then they're less likely to spend their money with you, and your potential revenue is capped.
104.
by dado3212 4y ago
Can also attest to them having comfy t-shirts if you're interested in that option.
105.
by dado3212 4y ago
As someone who’s run literally hundreds of A/B tests, many of them on the backs of UX research with users in the field, people have no idea what they want. The anecdata is a place to investigate, but never the end of the journey.
106.
by dado3212 4y ago
Agreed. It’s not necessarily 80/20, it’s just that power laws show up a LOT. 90/10, 99/1, etc.
107.
by dado3212 4y ago
I’m so confused by this argument. Why do people think that gait and biometrics data is so valuable for running ads? Trust me, there are millions more features that are more closely predictive that top ML models are just leaving on the table
108.
by dado3212 4y ago
> I was used to dealing with obfuscated code from Minecraft I wonder how common this is for engineers. I’ve been in industry for the last 4-5 years or so, and it was how I got a pretty big jump into programming. Talked to a few similarly
109.
by dado3212 4y ago
This isn't "I can integrate WhatsApp with Slack" this is "I can set my default text app/call app/video app from iMessage/Phone/FaceTime to a third-party app". Same with setting default browser.
110.
by dado3212 4y ago
Also, you're not allowed to > preference the products, services, or lines of business of the covered platform operator over those of another business user on the covered platform in a manner that would materially harm competition I
111.
by dado3212 4y ago
AICOA feels wildly vague to me. It would make it unlawful to > materially restrict, impede, or unreasonably delay the capacity of a business user to access or interoperate with the same platform, operating system, or hardware or software
112.
by dado3212 4y ago
I think he's mostly right here in identifying the problem. I'd agree with the linked comments on this post that the solution (and AICOA as discussed by the link for the post) is likely bad/inadequate.
113.
by dado3212 4y ago
The two driving forces for higher costs are subsidization of demand (gov't assistance + insurance, all of which have some pass-through effect on raising the cost of services), and artificial restriction of supply (it's hard to cre
114.
by dado3212 4y ago
The law forbids that, saying that > A social media platform may not censor a user, a user’s expression, or a user’s ability to receive the expression of another person based on…a user’s geographic location in this state or any part of th
115.
by dado3212 4y ago
I don’t think that’s true, that’s based on one LinkedIn post that had some extenuating visa issues. Also hiring is still open for E5+ ML and E6+ generally.
116.
by dado3212 4y ago
They could maybe use like activity in addition to just tweets? Inherently though this system is going to be less accurate than the dataset that Twitter has access to. If a large chunk of users only engage in Twitter through DMs then an exte
117.
by dado3212 4y ago
Comments on HN or Reddit or many other sites with voting tend to be critical. If you agree, you’d upvote. Commenting “this” or “so true” isn’t useful, and doesn’t happen as much as people going “this is totally wrong”. Just selection bias.
118.
Zuckerberg Sees Project Cambria ‘Eventually Replacing Your Laptop’(uploadvr.com)
11 points by dado3212 4y ago|14 comments
119.
by dado3212 4y ago
How does this work?
120.
by dado3212 5y ago
I think (and this is heavy arm chairing so maybe good to back up with some due diligence) that Carpenter v. United States made it so this required a warrant. I think if you have a warrant though, Google will give over this kind of info. [1]

More ›