Enough vaporware, Creality

AI Lidar anomaly… what now?

With the K-series of printers being Creality’s “flagship” line, I can’t blame them for trying to go feature-for-feature with Bambu Labs. Overall, it’s hard to argue with Creality’s strategy, as they are selling a lot of printers.

Sure enough, despite some teething problems, it’s safe to label the K1 printers as a success, overall. For me, I skipped most of that by getting a K1C, but I’ve seen an early K1, the differences are pretty obvious.

Considering the rush to get to market, it’s easy to have delivery errors. At worst, you have bad hardware, or exploding glass doors. At best, you have software/firmware issues that are caught fast, and patched quickly. The input shaping bug is an example of something that lands a bit in the middle… caught early, but fixed late.

…But what about features that aren’t fully baked - and may never be? Take Full Self Driving on a Tesla… it’s sold as a feature, but it’s not that functional - and it’s not guaranteed to get any closer to complete. The analogue here would be the AI features on the K1s.

I frequently get AI error messages from my K1C. Most of the time it signifies nothing, and whatever caused the messaging condition doesn’t have any perceptible impact on the part print. That the printer just continues with whatever it’s doing, is a tacit admission that this feature DOES NOT WORK. The rest of the AI features do not detect a bad layer, a lack of material where there should be material, items that have lost adhesion, or even a workspace full of filament spaghetti. Regardless of whatever Creality intended for the feature to do, it’s broken.

What I don’t get, though, is not being honest about it. If you wanted your features to match up well with the X1C, but you fell short, SAY SO. Change the firmware and Creality Print HW monitoring to prompt you upon completion if the print was successful or not (instead of assuming it was, just because the g-code ran to the end). If you say that the print failed, prompt the user to select one or more reasons for the failure (tangled filament, 1st layer, clog, etc.). With that, you should at least be able to train your model better for the future.

Maybe then you’ll get it right in future printers.