dev1n’s claim is great, but what i heard it didn’t do well.
and time has gone.
i bet cursor 1.0 could be better than then dev1n now? thanks
Claude code is what dev1n promised. Cursor is close.
Yeah I somehow can’t get my head around the CLI user interface on Claude Code.
Any news on dev1n? Saw some websites posting more critical reviews of the inital claims.
I just started researching claude code and came across this video of how to use it in cursor. I haven’t tried it but it looks like you can set it up so it looks like an agent window. I am going to try it later today but if you do, please let me know what you think.
Here is the video on how to do it: https://youtu.be/u-GGktUnSgg?t=179
1 Like
how they differ? i only know claude code charge you token by token.
thanks
Claude code is very expensive. A workaround is to buy their $200/m Max (or sth) subscription.
There is also Manus - a chinese Dev1n.
Considsering how much you can get done with Cursor for $200, yes I was also surprised how fast Claude code guzzles up tokens.
How well is Manus working?
Not sure, it got a lot of praise once it was out. Maybe it just got run down by Claude Code.
My gut feeling is they all are kind of similar. Cursor is the GOAT on
- interconnected modelling (Tabtab knows what you chatted, chat knows your manual edits, etc.)
- pricing (1 token for up to 25 tool calls)
- and a little bit behind on token expensive parallel background agents.
They are very quick on racing back to top. E.g. when Stormsurf had the edge on agent use.
Unfortunately their design guy does not think like a programmer, his stuff is often in the way of the flow.
1 Like
@T1000 edited quite some stuff in my last post 
1 Like
Ah, sometimes designers get in the way, and sometimes programmers do too 
Might need an systems architect instead lol.
In my personal opinion, cursor is currently the most user-friendly for AI programming, at least among the IDEs available in China