Product is really good, especially the memory feature. I can actually feel it learning over time and that is impressive. However, the UI and UX need a lot of improvement. Task summaries are not very readable and edited files are not viewable, which makes the experience a bit frustrating.
That said, you guys did a great job overall. I would strongly suggest giving it the ARC AGI2 Benchmark test because honestly, from my usage, it is crushing other tools like Cursor and Claude Code. I truly believe it will surpass the existing systems out there.
For reference, I compared Claude 4.5 inside Claude Code with the Claude 4.5 that exists here and this version really outperformed it.
Just focus on building a better UI and UX because trust me, this will absolutely hit.
Report
59 views