The main thing why I started using it too late was the slop. it's in the end AI generated, yes it can take your tasks but i never felt in my personal use case how it can help me when it's just generating slop. I used claude cowork more than openclaw after trying once that too in a cloud container since I'm afraid of its security

GitHub CEO also raised 60M for 'entire' to bring agent context to git. The dust is yet to settle here as it's difficult to bring a paridgm shift from today's git workflows

building because its always the dopamine from the coding agents than the problem getting solved. Github contribution graph is rigged because higher number of commits doesnt make you a better engineer. We needed this blog, ty

yep joining here late but docker images are not affected as we saw on twitter

afaik the trivy attack was first in the news on March 19th for the github actions and for docker images it was on March 23rd

I'd be interested to know how this verification is happening, user transparency would be helpful

isn't a similar thing done by entire cli? the startup which raised $60M seed recently

same thought, we've been using smartphones since 2 decades now, and not just we dont have a problem with qwerty, but anything new will be requiring more cycles to get accustomed to

I can touch-type with flow typing in GBoard at this point.

That's to say, I'm writing this comment on my Android phone without looking at the keyboard.

QWERTY is in my muscle memory in such a way that words have become writable as single stroke characters.

I really, really doubt this Keybee thing can be an improvement over that in any way.

I really want to see the visualization of words as the swipe typing patterns. I tried doing it on paper and realized I couldn't understand it just by looking but once I started visualizing it and swiping in my head I could start to get a feel for it. The tricky part is figuring out where in the keyboard the stroke begins

I also would like to implement all those texting algorithms but I need the help of the open source community at this moment. Cheers!.

People riding horse buggies probably thought the same when powered vehicles first came about, and look at the world now. You won't know unless you give it a honest try.

QWERTY won't be replaced on phones until there is a full phase change in how people interact with their phones that absolves us of keyboards entirely. Anybody here who thinks otherwise is welcome to make an offer to buy my two decades of notes on the topic.

>People riding horse buggies probably thought the same when powered vehicles first came about, and look at the world now. You won't know unless you give it a honest try.

This is a ridiculous non-analogy.

I'm flying a jet airplane, and you're telling me to give Ford Model T a try because you don't understand flight as a concept.

Or, in this case, Flow typing.

From Keybee's website:

>Some syllables and some words can be inserted through a simple combination of tap & swipe (we call it twipe) greatly reducing the number of touches for typing a text. For now the twipe is limited to the adjacent keys. Keybee Keyboard is swipe friendly.

I am typing an entire word with one "twipe" on GBoard.

Each word.

I'm done with touchscreen input methods that require me to think about tapping letters. I don't think in individual characters, and I don't type in them either.

Let me know if I can make it any clearer.

Also gboard is the best keyboard for that. Nothing else implements a prediction model over a number of words as far as I can tell. Or if they do, they fail really badly at it.

Swipe keyboard on Microsoft Lumias was better.

I think Android is only catching up to it in the past 2-3 years.

Sadly, Lumias went the way of the dodo, and I don't have a need for that sort of input on something that's not a phone.

Whatever Microsoft put out as a keyboard app for Android is different, they didn't implement the same UX.

Out of the swipe keyboards I tried for Android, GBoard worked the best.

Weird thing to say, why would a random port be peak efficiency that can’t be improved?

awesome, does it also show if its stuck somewhere like auth or requires human in the loop?

meanwhile me adding to the .md file, "you're not a LLM and instead a super intelligent agent to help me find HN trends"