drew

foresight prevents blindness

[CJ☔️ @samwakinsvv]

This video would have been generational if he had lost

[The BMJ @bmj_latest]

Hantavirus outbreak should reset WHO’s default approach to airborne risk
https://t.co/LHfwL4fFRw

[𝒸𝒽𝓇𝒾𝓈 )|( @drownedblond]

música mais importante da história

[drew @bxyonddrew]

security at the bar called the office in davenport doesn’t wash his hands after he pees and he’s touching everyone’s IDs

[Tom Barton @tjbarton83]

New report finds links between Iowa’s high cancer rates and environmental exposures — like pesticides, nitrates, PFAS and radon — raising concerns about Iowa’s intensive agriculture footprint and the state's increasing cancer rates. https://t.co/hTargIcMJU

[Education Week @educationweek]

Teachers Say Behavior Problems Aren't Just About Students. It’s the Parents https://t.co/JdyoyigLkJ

[Mike Trautmann @DMRPoliticsCzar]

Just how much has @IAGovernor's education savings accounts reshaped K-12 education? 99% of Iowa's private school students are now publicly funded, costing an estimated $350M next year. https://t.co/YrKmDrRLJK via @HelloTimWebber @svhernandez

[BURKOV @burkov]

LLMs process text from left to right — each token can only look back at what came before it, never forward. This means that when you write a long prompt with context at the beginning and a question at the end, the model answers the question having "seen" the context, but the context tokens were generated without any awareness of what question was coming. This asymmetry is a basic structural property of how these models work.

The paper asks what happens if you just send the prompt twice in a row, so that every part of the input gets a second pass where it can attend to every other part. The answer is that accuracy goes up across seven different benchmarks and seven different models (from the Gemini, ChatGPT, Claude, and DeepSeek series of LLMs), with no increase in the length of the model's output and no meaningful increase in response time — because processing the input is done in parallel by the hardware anyway.

There are no new losses to compute, no finetuning, no clever prompt engineering beyond the repetition itself.

The gap between this technique and doing nothing is sometimes small, sometimes large (one model went from 21% to 97% on a task involving finding a name in a list). If you are thinking about how to get better results from these models without paying for longer outputs or slower responses, that's a fairly concrete and low-effort finding.

Read with AI tutor: https://t.co/MipHHO6rjX

Get the PDF: https://t.co/XQrqiaGwIO

[Joshua McElwee @joshjmcelwee]

Pope Leo said in a Christmas Eve sermon that the story of Jesus being born in a stable because there was no room at an inn should remind Christians that refusing to help the poor and strangers today is tantamount to rejecting God himself. For @Reuters

https://t.co/Z91z9fGMKB

[Coach Joe Stokowski @CoachStokowski]

Football is evolving faster than ligament and tendon strength.

[*“Volk” intensifies* @TinctureDrone]

when uber driver speeds through a yellow light rather than stopping

[David Lingenfelter, PhD @dlingenfelter]

A Swiss study found that sharing classrooms drives viral transmission significantly more than brief close contact.

Consequently, improving ventilation is likely more effective than solely reducing proximity.
https://t.co/z9IGQC3nkL

[IL Politihub @ILPolitihub]

Chicago hospitals funnel patients into long-lasting guardianships, angering friends and family https://t.co/ClZhfgdPX9

[drew @bxyonddrew]

RT @matthewstoller: Private equity giant Sycamore partners, in their first move after buying Walgreens, is denying paid time off for six ma…

[Noah Garfinkel @NoahGarfinkel]

EXPIRED MILK SAYS MOLDY BREAD IS OKAY, SHOULD BE EATEN

[Ford Fischer @FordFischer]

Police officers handcuffed what they believed to be a 16 year old student armed with a gun at his high school based on an AI system telling them so.

It turned out the student had a Doritos bag, not a firearm, and the AI may have mistaken the two.

https://t.co/JCLKuAXqc1