Next-generation AI models

Re:Next-generation AI models by codetrotter on Tuesday June 21, 2022 @10:30PM (#31831431)

I use Tabnine completion in JetBrains CLion when I write Rust code. I’ve found it to be quite useful. Haven’t tried GitHub Copilot yet, so don’t know how Copilot compares to it.

[ Parent | on HN ]

Re:Next-generation AI models by speedgoose on Wednesday June 22, 2022 @12:36AM (#31832062)

I used Tabnine before GitHub Copilot and I never looked back. It feels a bit like upgrading from GPT2 to GPT3.

[ Parent | on HN ]

Re:Next-generation AI models by snorremd on Wednesday June 22, 2022 @01:37AM (#31832418)

I mean, that's literally what happened. At least initially Tabnine was based on the GPT2 model trained on code. Then GitHub launched Copilot using the OpenAI Codex model which is based on GPT3. I guess this is why several people have commented on the marked improvement when adopting GitHub Copilot.

I have no idea how Tabnine builds their models today, and how they perform compared to Copilot. I guess one advantage could be lower latency in suggestions that they claim come from training smaller more specific models. But the way I find Copilot working for me is that it thinks about the same time I need to think and then it makes a suggestion for a good chunk of code. If my thinking and Copilot's thinking match up then I can save myself a good bit of typing.

[ Parent | on HN ]

Re:Next-generation AI models by codemanjoe on Wednesday June 22, 2022 @02:35AM (#31832701)

There's a decent comparison outlining the differences by Tabnine's CEO here (disclaimer - I work there). https://twitter.com/drorwe/status/1539293063117516801

[ Parent | on HN ]