X

[X @al0x99]

working with local authorities to train an vision model to recognize asbestos pattern from imagery, maily provided from esa. That would help me into speeding up the training. Doing that for free for them, that would help schools and hospitals since in italy it's a major problem. I can show proof on what i am doing :)

[X @al0x99]

@METAPCs 1990,90

[X @al0x99]

@Mr_Salio This fcker its a scam. report @mr_salio shitpostwe

[X @al0x99]

@elshayib_ Try using muse 1.3 :)

[X @al0x99]

@Ananth7e model it's not the best but not bad. Still, lower than muse 1.3 max you can easily feel it

[X @al0x99]

@daradoescode Please send the prompt, curious i want to test the same

[X @al0x99]

on supergrok heavy. At xhigh + fast mode, for me it seems pretty good actually. Worked for 3 hours on 3 projects at a time and i'm at 7% usage, much faster and better than grok 4.6 without any doubt. I consume 15% of astra each week to analyze my prompts, outputs and everything it can so it optimizes my flow, do the same and you'll see how your workflow will improve.

[X @al0x99]

@pancakeufo shit

[X @al0x99]

same thing, 2 days ago i got the renewal. That kind of behavious is disgusting. It's even worse than seeing same model downgrade day by day. In the renewal pay you should have an overview like that:

- Features you won't have with this renewal:
-- only 100$ credits for other api usage.

Then you decide. I don't blame them, everyone knows the crisis with the chips but, hey fck off.

[X @al0x99]

it's better than luna, 100%. For me it's perfoming 90% i can say of 5.6 Sol Ultra honestly. It worked for 10 hours without my intervention (gave it very good instructions), and on contributor plan it consumed NOTHING. Was nothing so importand, but it build independent pieces for my software that were really really good, even Astra told it was really good. Got the 45€ plan of Muse Spark and enjoying it till it lasts

[X @al0x99]

@aapakari Totally agree! Muse is so good and cheap. Imo better than sol.

[X @al0x99]

@Mr_Salio please report this guy. Everything he says is fake

[X @al0x99]

@Mr_Salio all fake

[X @al0x99]

@0x0SojalSec go fck urself and those fake post

[X @al0x99]

Nothing too crazy but if you have a projects or something and need some cpu or ram, i would borrow you some. Can't guarantee anything. Also have 100TB+, and a dedicated 2.5gbps line. No illegal shit please.

[X @al0x99]

@KSimilien @theo @TheAhmadOsman it's a frontier model. performing better thant 5.6 sol for me. easily

[X @al0x99]

@mackinleysmith @maria_rcks dont ever open any pr. She will copy your work, merge it and never mention you. Just ask any ai to look into what i am saying and you'll find a pattern. I think this is fcked up.

[X @al0x99]

this is how open source works?

you send the PR.
they don't merge you.
a trusted collaborator self-merges a wider version of the same feature, with no mention of you.
five minutes later a maintainer closes yours as "leftover hygiene."

the feature is on main. you are not in the history.

that is not a missed citation. that is the method. merge theirs first, then retire yours, so the public record says they shipped it and you were leftover.

today, pingdotgg/t3code, three times, same afternoon, UTC.

16:58 maria-rcks self-merges #11519. no mention of #10796, which I opened Sep 8.
17:03 juliusmarminge closes mine as superseded.
I already had ACP command catalogs and native slash parsing. they shipped that Cursor path plus OpenCode/Grok. still no hash.

18:32 maria-rcks self-merges #11784. no mention of #11623 (MONKE2525E, Sep 13).
18:36 closed as superseded.

19:42 maria-rcks self-merges #12115. no mention of #12081.
19:44 closed as leftover hygiene.
peachesandcream118 opened #12081 today at 12:34. the overlapping trusted PR opened at 17:45. five hours. then merge. then hygiene.

if this were sloppy it happens once and you still write the #. it happened three times in three hours and the # never appeared.

they already know the honest version. #9507 says "#1732 by @Aditya190803". #8803 says "supersedes #7627". today they chose the version with no name.

https://t.co/sHKDl3P4gF
https://t.co/ItUVbjWgvF
https://t.co/DVAYJ1crPS
https://t.co/nP4x7Q6phk
https://t.co/lSIDEAGqhC
https://t.co/5y5h3erKhb

[X @al0x99]

@maria_rcks yep, so we can copy paste code from other maintainers exactly how you do it 😊 theo keeps talking about your talent but i mostly see copy paste and push, a i wrong here? Open for discussion