BrandonJung

8 ms·

31.
by BrandonJung 4y ago
it is possible that on a long snippet that you could be copying code from non-permissive code. In addition GitHub themselves have said that this is not resolved and will likely need to be adjudicated in the courts. The only way to avoid t
32.
by BrandonJung 4y ago
copilot actually pushed Kite out of business but I am here at Tabnine and we have been doing great. MSFT is always tough to compete with but I did it at GitLab before and I think with our strong take on personalized models, ability to run
33.
by BrandonJung 4y ago
what about those that have code on GitHub that is source available but not licensed for reuse?
34.
by BrandonJung 4y ago
they used all the code in GitHub regardless of license. hope they avoided the Oracle code ;-)
35.
by BrandonJung 4y ago
one way to minimize this is to train on your own trusted code. You do need a reasonable amount of good code (ideally with good comments too) this is one of the options that we have here at Tabnine. Train on your GitLab, Bitbucket or GitHu
36.
by BrandonJung 4y ago
Come try Tabnine and if you need a custom model for your small OSS project please let me know.
37.
by BrandonJung 4y ago
please take a look and also note that Tabnine while being more secure has also continued to evolve. We also have a free option that we have stood behind for 3+ years. https://www.tabnine.com/tabnine-vs-github-copilot
38.
by BrandonJung 4y ago
Tabnine is a competitor! We have been doing this for 5+ years with more developers than Copilot. Please take a look at the posts and if curious visit https://www.tabnine.com/tabnine-vs-github-copilot
39.
by BrandonJung 4y ago
No they did not. You have to train on only fully permissive code to ensure that is not a problem
40.
by BrandonJung 4y ago
Tabnine has been working in this space for more than 5 years and we would concur with much of the sentiment here on the importance of being able to adjust the length of the suggestions and ensuring the model is trained on ONLY fully permiss
41.
by BrandonJung 4y ago
would you prefer it trained ONLY on your code or are you ok with the broad use of non-permissive code used for CODEX?
42.
by BrandonJung 4y ago
when we first built Tabnine we had confidence percentages next to suggestions. Do you think this would help?
43.
by BrandonJung 4y ago
this is consistent with the feedback from many of the users we have talked to as well (transparently I am with Tabnine). Long blocks of code are difficult to digest while short quick ones can be very quick AND easy to validate the logic.
44.
by BrandonJung 4y ago
Licensing is a critical question that is often not considered. Code trained on non-permissive code (think Oracle API's) has very significant risk, ask Google. We took a different tact three years ago in building Tabnine BUT went with
45.
by BrandonJung 4y ago
As Copilot is becoming generally available, this might be a good time to write a comprehensive comparison between the two leading AI assistants for software development Tabnine and Copilot by Microsoft. Details here are from our CEO and Fo
46.
by BrandonJung 5y ago
Looks like a great way to markup compute 100%. Good margin for Azure.
47.
by BrandonJung 5y ago
here is the link for details. https://github.com/codota/TabNine/issues/432
48.
by BrandonJung 5y ago
quick response here for your request... your request for low opacity text is here. Also your concern about licensing is never an issue with Tabnine as we only train on fully permissive code.
49.
by BrandonJung 5y ago
for those of you like this concept but 1. did not get into the alpha (or want something that has lots of great reviews) 2. need to run locally (security or connectivity) 3. want to use any IDE... please try Tabnine