Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
master_yoda_1
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
13 ms
·
91.
▲
by
master_yoda_1
6y ago
yeh don't f k individual privacy for ads, and let a sh t company like FB make billions of profit
92.
▲
by
master_yoda_1
6y ago
Looks like everyone at hackernews have intelligence of a five year old :) Sad to see for so many human being their brain never evolve.
93.
▲
by
master_yoda_1
6y ago
This is everywhere. I have to deal with lots of jokers at work, who just suggest to use GPT-3 without understanding what it is :)
94.
▲
by
master_yoda_1
6y ago
why don't use TF-IDF and use just elastic search?
95.
▲
by
master_yoda_1
6y ago
How you feel if somebody claim they know java just because they can call elastic search api (or put any tool written in java). Are you going to hire them and put them on a project which involve coding in java?? the issue is false claim on r
96.
▲
by
master_yoda_1
6y ago
I agree with this. Changes for Deep learning models is so fast there is no point maintaining a super reusable code.
97.
▲
by
master_yoda_1
6y ago
Did you understand the correlation? And what part of the source code is doing it? Or you had just fun without any understanding?
98.
▲
by
master_yoda_1
6y ago
Nice but i am scared people write in their resume that they train a GPT model from scratch. And when asked in detail they will accept just ran minGPT without understanding it. This is the AI story now a days. Best solution is to ignore minG
99.
▲
by
master_yoda_1
6y ago
It can easily do copy paste programing, in the interpreter languages.
100.
▲
by
master_yoda_1
6y ago
I think it will filter copy paste programmers, and that is for good.
101.
▲
by
master_yoda_1
6y ago
Yeh then advertise it as a tool rather than AI. The problem is that sorkle is trying to sell snake oil on the name of Stanford and AI. Under the hood it is just a data generation pipeline. Remember you can't put label on random data. S
102.
▲
by
master_yoda_1
6y ago
is not this done for years and called synthetic data generation, simulation etc.
103.
▲
by
master_yoda_1
6y ago
Looks like they selling snake oil on the name of Stanford and AI :)
104.
▲
by
master_yoda_1
6y ago
I think this is known for years. Is't it?
105.
▲
by
master_yoda_1
6y ago
Yeh I experienced the same
106.
▲
by
master_yoda_1
6y ago
Initially I lost weight but hit a plateau later and got demotivated
107.
▲
by
master_yoda_1
6y ago
Going to gym, eating salad, skipping lunch
108.
▲
by
master_yoda_1
6y ago
“How to get 6 packs abs”, trying to solve this for at least last 10 years but it’s a really hard problem.
109.
▲
by
master_yoda_1
6y ago
Looks like I need to start learning julia
110.
▲
by
master_yoda_1
6y ago
I am confused about the FAST part, it is faster than all the other implementation (some of them are in c++) or it is just julia implementation and you think it is fast? I am asking because if julia is faster than c++ for ml/dl I would
111.
▲
by
master_yoda_1
6y ago
One very dangerous thing about university I notice is that it encourage age discrimination. And that propagate to the industry hiring. Very few people with family will choose to go to university full time, and hence can't further their
112.
▲
by
master_yoda_1
6y ago
Forget machine learning or ai and learn a compiled programming language in and out. It will help in long run.
113.
▲
by
master_yoda_1
6y ago
By the way this is not a project. Please don’t call it a side project you are misleading people.
114.
▲
by
master_yoda_1
6y ago
what about the looting on the street?
115.
▲
by
master_yoda_1
6y ago
respect
116.
▲
by
master_yoda_1
6y ago
I think we should not compare if anybody is capable or not. Here the most of the work is done by azure engineers and none of them are mention in the paper. So no even open ai can’t do it without azure infrastructure.
117.
▲
by
master_yoda_1
6y ago
Yes then credit to azure engineers not open ai people
118.
▲
by
master_yoda_1
6y ago
the title is misleading
119.
▲
by
master_yoda_1
6y ago
One question: Why put more sh!!t on top of huggingface already really good code?
120.
▲
by
master_yoda_1
6y ago
If it takes 6 month to come to speed for an experienced engineer, I would say you are working with dinosaurs, in a jurassic age company.
More ›