larwent

3 ms·

1.
by larwent 2y ago
I’ve been using something similar called Twinny. It’s a vscode extension that connects to an ollama locally hosted LLM of your choice and works like CoPilot. It’s an extra step to install Ollama, so not as plugnplay as tfa but the license i
2.
by larwent 4y ago
Ask and you shall receive https://i.imgur.com/fN5EoIB.png
3.
by larwent 4y ago
The exception to this appears to be "Show HN" posts. E.g. https://news.ycombinator.com/item?id=31516108
4.
by larwent 4y ago
> Several high profile companies have already announced layoffs in recent weeks Which companies are these exactly? What do long tenured developers see in this latest new cycle? Are Bloomberg just stirring up drama is or there a strong po
5.
by larwent 4y ago
Seems pretty simple. When we create upper and lower boundaries to some score, people with lower scores have more space to overestimate and those with higher scores more space to underestimate, causing the perceived score to trend towards th
6.
by larwent 4y ago
Unless I missed something, this article doesn't explain WHY random data can result in a Dunning-Kruger effect. The relationship between the "actual" and "perceived" score is a product of bounding the scores to 0-100
7.
by larwent 5y ago
It's the typical GANS face layout, with a blurry background, eyes centered and cropped to the face. It's certainly possible those are could be real people, but in my experience law firms usually have upper-body shots of the lawy
8.
by larwent 5y ago
The lawyer profile faces instantly give the site a fake vibe as they look like those generated by GANS AI.