#statstab #554 Does Regression Produce Representative Estimates of Causal Effects?

Thoughts: Regression creates an “effective sample” that may be unrepresentative of the pop of interest even w/ a representative sample

#ate #causalinference #regression #sample #weights #representative #observational #inference #LATE

https://onlinelibrary.wiley.com/doi/abs/10.1111/ajps.12185

patreon sketch for the massive @hungrylizrdaftrdrk.bsky.social who's keeping well fed for those gains >:3 

If you want to get a monthly sketch and other goodies check out my patreon.com/Toomuchgoo

#lizard #reptile #sizedifference #gym #muscle #gut #bulges #multiple #prey #struggling #weights #pecs #belly

"They're made out of weights"

https://maxleiter.com/blog/weights

#HackerNews #weights #fitness #innovation #technology #blog

more ai on open source ai, did I prompt it to say this, maybe? #tokens distilled #smb ai #american big tech ai fever dream #snake river canyon jump #don't be evil knievil #the big short 2 #ipo #valuations #p/e ratios #weights are math #source:deepseek
"
Where does open source AI fall into place?

Open-source AI is the escape hatch from the entire sovereignty vs. subsidy cage match.

It makes your predicted "free trade zones" redundant, because the whole internet becomes a free trade zone.

It undermines US big tech's valuation premium (good for competition, bad for investors).

It partially neutralizes China's data-sovereignty advantage (if you run the model locally, you don't send them data).

It cannot be stopped without breaking the internet or banning math.

The only thing open-source can't solve is the compute gap for training the very largest models. So governments will fight over chips and training clusters, but they've already lost the battle over model distribution.

That's the real shakeout: Closed-source frontier models as state assets, open-source everything else as a global commons."

Build it, and it will come.

I don’t think I’m being too boastful if I say I’ve always been ‘OK’ at whatever sport I’ve competed. This though was more down to my fitness level supplementing my limited technical skill. Now, that’s one thing I am happy to brag about.

For a s long as I can remember, being ‘fit’ has been possibly my top priority in life. I’ve always taken the view that I’m a runner, not a fighter! If someone wants to give me a doing, they’ll have to catch me first.

I remember stepping off a train in Glasgow in the early ’80s. I was wearing gear most associated with the Rockabilly / Ted revival of the time. I was ‘greeted’ by a group of punks. Punks and Teds didn’t get along, it’s fair to say, and though I’d say I stood on the periphery of both groups, those guys weren’t really in the mood for general discussion.

They ran at me. But I knew they had no chance. I was cocky and confident enough to gesticulate a parting message as I disappeared up Buchanan Street, into the gloomy, autumn night.

Anyway, I digress.

Even now, at my ripe old age, I still work out to some degree every day. On top of regular pickleball sessions /matches. That said, since my cardiac arrest three years ago, I promised my family I wouldn’t push myself too hard.

And that means I now have a more regular, but moderated fitness regime.

For those wishing to embark on that journey, those determined to get off the sofa and feel better about themselves … what’s stopping you? You don’t need to set any athletic records. You don’t need to look good.

Start slow. Build up pace and or stamina with incremental weekly targets. The beauty about keeping yourself fit is you’re not competing against anyone. Don’t be too hard on yourself. If you don’t see continuous improvement, don’t sweat. Well – maybe sweat a little bit more … but you know what I’m saying.😉

One you start, give it a week or so. Push through the discomfort you are likely to feel. But I guarantee, stick with it, and you will notice a difference to your weight / stamina / general fitness very quickly.

And when you start to see that improvement, your motivation levels will rise. You’ll be hooked!

‘Fitness’ doesn’t come easy. It needs a bit of encouragement. It likes routine.

Build it, and it will come.

I promise.

Rope Physics Toolkit #Ropephysics #Unity #Mobile #Performance #Ropesimulation #Pulleys #Weights #Collisiondetection #Ropeconnectiontypes #Sceneviewhandles #Burstcompiler #Unityjobs #AssetStore

https://u3dn.com/packages/rope-physics-toolkit-154662

If you are running Google Chrome on Windows and wish to disable the illicitly downloaded 4 GByte AI model, here are the steps to go though:

Deleting the file (weights.bin) is often ineffective as Chrome will automatically re-download it upon relaunch unless the associated AI features are disabled.

Standard Settings: In Chrome, Go to Settings > System and toggle off On-device GenAI (or "On-device AI") Experimental Flags:

Enter chrome://flags in your Chrome address bar and disable the following:

#optimization-guide-on-device-model
#prompt-api-for-gemini-nano

After disabling the flags, you can manually delete the storage-heavy folder to reclaim space (roughly 4GB).

Press Win + R and paste:

%LOCALAPPDATA%\Google\Chrome\User Data\Open your profile folder (usually named Default).

Find and delete the OptGuideOnDeviceModel folder.

As an experiment I have deleted just the existing weights.bin file and re-created a read-only 0 byte version. I can then monitor this to see if it's changed in the future.

Hopefully this will disable these illicit AI features and prevent it from being re-enabled in future.

#Windows #Chrome #Google #Weights.bin #AI #LLM

Context Is Software, Weights Are Hardware

https://www.aravindjayendran.com/writing/context-is-not-learning

#HackerNews #Context #Is #Software #Weights #Are #Hardware #Technology #AI #Learning

Should runners lift heavier weight, lower reps or lighter weight, higher reps? A strength coach explains all

https://quokk.au/c/running/p/848335/should-runners-lift-heavier-weight-lower-reps-or-lighter-weight-higher-reps-a-strength-coach