I need to make a choice while filming the next #heliboard video (which I am starting very soon)

The script basically only calls for phone screen recordings in portrait and shots of me talking. The last video used choice 1, but it had other shots too.

How should I go about this one do you think?

Feel free to boost!

Neither of the options are great in my opinion. Landscape phone screen recording won't work for these shots unfortunately... trust me, very disorienting.

Got 576 today with the left hand.

Check out the total gathered caught the odometer ticking over. 😎
#heliboard

I just made myself a mastodon list for #heliboard users. If you follow me, and we aren't mutuals, reply or send me a message. I will follow you and put you in there so I can keep tabs on what people are up to!

I am hoping it will help me be stay motivated to open it and see people that would be writing toots using my code when we get there. 🙂

Tangible achievable goals.

Anyone interested in helping me out by editing a video for me? I have to make a video about #heliboard background gathering, but that isn't part of the nlnet grant, so I can't be paid for it and it will bring me closer to the wire on getting the agreed upon work done.

It's kind of important that there are detailed instructions on how to use the privacy tools for controlling what background gathering actually collects. Don't want to roll it out unless we have that available for contributors.

Idk Heliboard is testing me on words like "zoological" and we just mixed up "hey" with "get"

It gets "put" and "out" focused civvies confused too

#Heliboard pwease

Mahbe inshiuld go back tk thumb typing bjt i dont loke that either

How.the fixk does anuome type anutionf kn a phkn??

Hey kid, wanna buy some graphs? 🥸

This is how the google library performed on batch 1 of #heliboard data.

I didn't have these graphs before now, just got the testing harness to spit them out. (Progress! Needs more testing though.)

Right now, they are CSVs I load into calc; I am going to have these, and 70 other graphs, generated by a python script soon.

perfect = top suggestion
good = on the suggestion bar
bad = long press for suggestion
fail = user had to force accept gesture

Got a base going for the #heliboard gesture testing harness.

According to my _not_ tested or verified outputs, the google library is getting gestures right pretty often.

I was ranking it this way:
correct word -> 1
on the suggestion bar -> 1/3
in the long press area -> 1/21

Then average...

batch-1-en-US, it got 0.88286 overall. The words we collected don't follow a typical word distribution; it isn't leaning on word frequency much. Gonna be work to beat that...

Now that we have so many submissions, and with a several months left in the gathering, I am considering changing the distribution of words we pick up in the active gathering in #heliboard

I am thinking of making it more closely resemble the real distribution of the words, and introducing a small bias towards short words.

We have a lot of really odd words, which is good, but we do need to make sure that we get more common words.

Let me know if you would find this more boring to contribute to!