despite having tweeted across many different topics, i noticed i didnt introduce myself at all.
i'm mehmet utku öztürk from ankara/türkiye. as of now, i am 18 years old. i've been working on software since i was 8, starting with game development&small apps. i've bounced between many different fields from gamedev to cybersecurity and enterprise software to deep ml research. i'm currently focused on ai and ml with a particular interest on neurosymbolic systems; with the dream&aim of contributing to the achievement of artificial consciousness. i am interested in building new stuff, lifting weights, longevity/biohacking, martial arts, psychology, sudoku, cold brew and good conversation.
you can check my work&articles on my personal website at https://t.co/MBCTevcuvY and reach me via email at [email protected] ; i'm not very active on social media. cheers 🥂
Everyone found out a fruit fly brain could launch a memecoin on Sept 10.
But @mutkuoz had already made a REAL fruit fly brain play DOOM on Aug 20 — before the connectome hype even started.
139,255 real neurons. 2.7M synapses. DOOM fed directly through the fly's visual system.
Nobody noticed.
So we launched $FLYDOOM on @socialhood.
Here's the part that matters. 100% of the creator fees from every $FLYDOOM trade are reserved for github:mutkuoz. Not for the launcher. Not for us. He doesn't need to know the coin exists: the ETH accrues on Robinhood Chain and waits.
When he wants it: drop a one-time code in his GitHub bio, verify, withdraw. No wallet needed to start earning. No permission needed from anyone.
robinhood:0x4eb990547bce4a982432ca88cf5fae7eed1a2d35 = the fly that launched a coin.
$FLYDOOM = the fly that played DOOM first.
Earlier date. Real repo. Real research. Real dev getting paid.
Repo: https://t.co/07FcRURW8F
@zmbnski@notselfmodel most of the time yes. in my opinion, using a frozen connectome & modifying the environment to better match with an environment that a fly connectome would be evolved for keeps the experiment more "authentic", which is what i did at flydoom
i wanted to thank everyone who contacted me for their support for flydoom. honestly, it started as a small fun project on my spare time and i did not expect this kind of interest or support. thank y'all <3
despite the market cap drop of $flydoom, i'll be continuing the development; trying the methodology with the new hhmi janelia dataset instead of princeton's 2023 fafb and finalizing&publishing the research paper.
i'm open to any kind of suggestions, ideas, questions, collaboration, help, and conversation in general. take care
If you told me 3 years ago I’d be spending $200/mo to spend my weekends with a model named Astra I would’ve thought I’d be up to something completely different