rgavuliak

11 ms·

91.
by rgavuliak 3y ago
Micropayments don't work business-model wise. There was a "national model" in Eastern Europe, that offered one subscription to multiple services, but the company evolved to provide individual solutions. While it was good for
92.
by rgavuliak 3y ago
It's interesting that a lot of the posts in this thread focus on the product outcomes of the SV culture, but ignore the point of the article, how the endless quest to squeeze out the 1 % marginal productivity creates this kind of outco
93.
by rgavuliak 3y ago
I am looking at the data analytics industry as a whole and being involved in communities of data practicioners. Most of these people use cloud DBs (Snowflake, BigQuery & co) since there's less dependencies on DB Admin type of work.
94.
by rgavuliak 3y ago
Do you realize most reporting & analytics use cases don't use SQLLite Databases?
95.
by rgavuliak 3y ago
How easy is it to clawback money donated to a politician through a 3rd party? I am not a lawyer and don't live in the US, but I imagine that being super difficult and lengthy as a process.
96.
by rgavuliak 3y ago
The distinction is scale at which OpenAI can make profit off of your work. Now this might sound trivial, but it's scale of fraud possible has been the biggest argument against online elections.
97.
by rgavuliak 3y ago
People say that this is just like any big IT corporation. It's not IT in Germany does a lot of things manually because it's been done like that historically and they can't fire those people due to unions.
98.
by rgavuliak 3y ago
I work for a larger start up (or a scale up) that is doing decently. Majority of the top leadership was promoted from within sometimes all the way from IC to VP and I think it shows. The org would definitely benefit from someone that would
99.
by rgavuliak 3y ago
They can work well for B2C, wouldn't expect the same for B2B businesses.
100.
by rgavuliak 3y ago
In online games and newspaper paywalls you have on average 1-3 % valuable users. Even if Apollo made up less than 1%, those are probably the valuable ones so it could be a good chunk of those.
101.
by rgavuliak 3y ago
More (and more recent) content will be need for further training of the models for them to stay competitive and up to date.
102.
by rgavuliak 3y ago
Don't you guys have phones?
103.
by rgavuliak 3y ago
The code snippet explicitly defines the index to be 0, 1, 3. It may just be a typo, but if even simple examples are written sloppily, how can I trust the sheet: df = pd.DataFrame( {'col_1': [11, 12, 13], 'col_2': [21, 22
104.
by rgavuliak 3y ago
Ok, in that case Python would be 0, 1, 3 and Julia 1, 2, 3. My point is that the example explicitly skips an index in the definition of a data frame for Python, but it doesn't for Julia.
105.
by rgavuliak 3y ago
The cheatsheet goes wrong already for the first example of declaring a df: - you could do a range in python (range(11, 14)) - columns are called col_1 & col_2 vs a & b (both sets are horrible names) - pandas defines index of 0, 1, 3
106.
by rgavuliak 3y ago
I have spent a good deal of time in the US during my previous job. To some extent I agree that the US has a lot of opportunity (not just work, but food, culture etc.) and it was great being there. At the same time I can walk the streets of
107.
by rgavuliak 3y ago
Because when buying alcohol, you can for the most part (outside of online purchases, but even then the courier can check an ID upon delivery hypothetially) verify it fairly easily. With online porn, the scale and complications doing it put
108.
by rgavuliak 3y ago
Almost feels like you could provide this to other merchants in some way/shape/form.
109.
by rgavuliak 4y ago
It's not even an issue of statistics - yes the traits might exist, and people might fit into it, but that doesn't mean it's predictive of any outcome related metric.
110.
by rgavuliak 4y ago
I train market mix model for high level attribution for US companies and it usually looks like TikTok doesn't have a strong impact on revenue compared to how much is spend on it. Might be that there is some reason that the model doesn&
111.
by rgavuliak 4y ago
I have done this quite a bit. I am originally a Lead Data Scientist that got a bit into product and recently I became an Engineering manager for a fully cross-functional team that also has Data Science in it. It works out great and I have n
112.
by rgavuliak 4y ago
Why this vs Looker Studio that is maintained by analyst and have version control for metric and dimension definitions?
113.
by rgavuliak 4y ago
I asked a similar question and was prompted to buy from Alibaba.
114.
by rgavuliak 4y ago
No, fy_iceworld had guns lying on the ground. The article mentions it too.
115.
by rgavuliak 4y ago
I was using a docker interpreter with Pycharm. It's fairly simple.
116.
by rgavuliak 4y ago
Lambdas now work with Docker
117.
by rgavuliak 4y ago
> why not run it on all the data you can Because more data requires more cleaning and standardization (with more edge cases). It also requires a bigger scale to obtain and process.
118.
by rgavuliak 4y ago
A blog like this only scratches the surface. A couple of simple examples don't really help decide on whether it's a good substitute.
119.
by rgavuliak 4y ago
At the same time for me Spam gets through that pretends to be tied to loyalty programs.
120.
by rgavuliak 4y ago
If their agenda is for people to live in a better environment than it's bad by default? I would understand your approach, if you'd be able to find the bias in the study or proof of them lying/manipulating facts, but isn'

More ›