News for hackers. Stuff that matters.

A comment on Qwen2.5-1M: Deploy your own Qwen with context length up to 1M tokens. Parent comment

Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by arkh on Monday January 27, 2025 @02:44AM (#42838433)
My hypothesis is code completion is not a text completion problem. More of a graph completion one.

So we may have got to a local maximum regarding code helpers with LLMs and we'll have to wait for some breakthrough in the AI field before we get something better.

[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by raincole on Monday January 27, 2025 @03:05AM (#42838590)
But these models don't work that well even for text when you gave them a huge context. They're reasonably good at summarization, but if you ask them to "continue the story" they will write very inconsistent things (eerily similar to what a sloppy human writer does, though.)
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by meiraleal on Monday January 27, 2025 @07:55AM (#42840633)
We should be able to provide 2 fields, context and prompt so the prompt gets higher priority and don't get mixed with the whole context.
[ Parent | on HN ]
Re:Qwen2.5-1M: Deploy your own Qwen with context length up t... by meiraleal on Monday January 27, 2025 @07:53AM (#42840622)
For this breakthrough to happen, big tech will need to hire software engineers again :)

But the good thing is that DeepSeek proved those breakthroughs are going to happen one way or another, fast.

[ Parent | on HN ]