Shifty_Schultz

[Shifty_Schultz @ASaltWeapons]

@jnyboy Disgusting. I demand trial by combat.

Midnight. You know where to find me.

[Shifty_Schultz @ASaltWeapons]

@punished_daniel Just do it as a third party certification company. Make a little "Green Data Certified" stamp and everything.

Don't even have to build the datacenters, they pay you a fee to inspect and you glaze them on tiktok for not obliterating water molecules out of existence.

Win-win.

[Shifty_Schultz @ASaltWeapons]

@yacineMTB Tough question. Which computer vision model do I need for this?

[Shifty_Schultz @ASaltWeapons]

@cremieuxrecueil Does this put him at better or worse odds of solving your problem than ChatGPT?

[Shifty_Schultz @ASaltWeapons]

@CarMarinkovic @cremieuxrecueil Didn't catch that, thanks.

[Shifty_Schultz @ASaltWeapons]

@HoffmanTactical Commitment issue tbh.

[Shifty_Schultz @ASaltWeapons]

@sever53590 @st_louis_stan Attending the same church is also one of the next highest predictors.

Iirc the winner in the age range category is between 25-30 years old, but meeting in higher education is it's own positive factor, so I'm not sure whether that's just reflecting college couples.

[Shifty_Schultz @ASaltWeapons]

@gaitanalyst buccal fat removal for the mind

[Shifty_Schultz @ASaltWeapons]

@eigenrobot The one drill sergeant that lectures trainees for 30 minutes on how applicable the behavior that he is observing is to the most recent chapter of his child psychology coursework while they do their pushups.

[Shifty_Schultz @ASaltWeapons]

@cremieuxrecueil This session is the anvil that pops the AI bubble

[Shifty_Schultz @ASaltWeapons]

@mfranz_on 90% of the time if you wanna find a model for something it's already on roboflow or github too.

[Shifty_Schultz @ASaltWeapons]

@eigenrobot If prioritizing extreme minorities at the cost of literally all else was just a matter of bad advertising, then I care even less to implement whatever your *actual* plan was.

[Shifty_Schultz @ASaltWeapons]

@333333RSJ @yacineMTB Similar. I guess deep-research is a prebuilt workflow that's run the same way as ultracode's custom workflows.

Just doesn't seem to have a reasonable agent limit. I told it to not use as many agents and it just didn't engage deep-research, so it may be built in.

[Shifty_Schultz @ASaltWeapons]

@firestormodk @yacineMTB Far better, a good third of the agents it deployed maxxed out their context windows and didn't return anything meaningful.

Iirc the prompt was something along the lines of asking if attention masking is useful in training image detection models reliably with less training data.

[Shifty_Schultz @ASaltWeapons]

@beaverd "A reaction drive's efficiency as a weapon is in direct proportion to its efficiency as a drive."

[Shifty_Schultz @ASaltWeapons]

@ShamashAran Meanwhile, when you ask 4.8 High

[Shifty_Schultz @ASaltWeapons]

@K0MODOS Ah yes, the JunkHeadRooms

[Shifty_Schultz @ASaltWeapons]

@eigenrobot coasting on aura alone