Hey everyone! Just saw the news that Tabnine has partnered with a major cloud provider (rumors point to AWS, but it's not officially named yet). This is a pretty big shift from their previous "bring your own model" and on-prem focus. 🤔
I've been using Tabnine Pro for about a year now, mainly for Python and TypeScript, and I really valued the privacy aspect. My big question is: **what does this mean for existing users, especially around data privacy and model control?**
Here are my immediate thoughts and concerns:
* **Potential for deeper IDE/cloud integration:** Could lead to features that directly leverage the cloud provider's services (like AWS CodeWhisperer's AWS-specific completions). Imagine Tabnine suggesting infrastructure-as-code snippets tailored to that cloud.
* **Data pipeline changes:** Where does our code go for processing now? Their blog says they'll maintain strict privacy, but the terms might change.
* **Pricing and tiers:** Could this partnership bundle Tabnine with the cloud provider's developer suite? Or maybe introduce a new, more expensive tier for the integrated features?
From a code quality perspective, I wonder if the underlying models will improve faster with more compute. I did a quick test yesterday comparing Tabnine's current inline suggestions with a few others. For example, writing a pytest fixture:
```python
# I typed: `def create_test_user(db_session):`
# Tabnine currently suggests a decent completion:
def create_test_user(db_session):
"""Fixture to create a test user."""
user = User(email='test@example.com')
db_session.add(user)
db_session.commit()
return user
```
Will a cloud-backed model make these suggestions more context-aware, like knowing I usually use factory patterns? Or will it just get better at suggesting AWS SDK boilerplate?
I'm cautiously optimistic, but I'd love to hear what others think. Are you worried about the pivot, or excited for new features? Let's discuss!
Happy coding!
Clean code, happy life