bryson

bryson

@brysonm.bsky.social

tell me the news. yelling at the telecom oligopoly for you. also giving the robots a chance. you're probably mad at capitalism, not technology 📍Toronto

661 followers1.2K following1.7K posts

serious offer: I will happily sort through your crazy tips if you're an overworked reporter

Molly White@molly.wiki

I've basically had to start ignoring tips that come in with clearly AI-written emails, even though I know I risk missing stuff. I just don't have the capacity.

despite the shockingly bad premise of the review, it does feel like it's correct about how much fielder and oppenheim make themselves characters betrays the original definition of documentary otoh, everyone is still in on the joke he's complaining about

This post is unavailable.

the most stressful part of this is these threats seem to be only mitigated through basic (yet somehow unlimited) incompetence of russian operations

This post is unavailable.

well sounds like a bit of a nightmare to vote in the Alberta separation referendum

Robson Fletcher@robsonfletcher.com

In the voting booth, it was a bit unusual dealing with 10 separate ballots. It was a process of unfolding, reading the verbose questions, then marking and refolding (each ballot needed to be folded twice) & keeping them all in order. Manageable but more cumbersome than a typical voting process.

bryson reposted

We've developed a new inference engine, called 'e'. It's the closest thing to alien technology which we have, and pushes 10,500 tok/s aggregate on a single DGX Station with the new @NVIDIAAI Nemotron 3.5 Super VL model. actual.inc/company/blog...

actual.incActual Computer › Case Study : GB300 and Nemotron 3.5 Super VLDepartment-scale AI from one NVIDIA DGX Station: Nemotron 3.5 Super VL at 8,000+ tokens per second, shared by your whole team through api.actual.inc.

lol plot twist: OP’s conclusion by the end of this thread turns out to to be that Swartz deserved what happened to him

This post is unavailable.

Imagine! Someone using artistic license on your words to score internet points. At least you’re still alive to block them

just in time for the AFU to start hitting data centres instead

Steve Herman@newsguy.bsky.social

President Trump, in a social media post, says Russia and Ukraine have agreed to an energy ceasefire "effective immediately" and warns Moscow and Kyiv: "Do not break it!"

also once you realize how APIs actually work, you figure out how much sit hidden behind a GUI

conputer dipshit@davidcrespo.bsky.social

“clicking around on websites that have existed for 8 years is a natural part of the human experience“ is one of the worst arguments against using AI I have ever heard, and that’s saying something. 99+% of the humans who ever lived have never done anything like it www.hottakes.space/p/a-slow-ced...

sometimes you don't in fact need to hand it to the terfs

Evan Urquhart@evanurquhart.bsky.social

I'm increasingly frustrated that people who are anti-AI aren't being taken seriously, and taking a page out of my thoughts on why the anti-trans backlash got as far as it did I'm starting to think the anti-AI side isn't extreme enough.

of course the golf courses were labeled in the legend

Toronto Archives@torontoarchives.cityof.toronto.ca

You get an expressway, and you get an expressway, and YOU get an expressway! In 1943 the Toronto Planning Board foresaw not only the Gardiner and the 401, but also superhighways along Bloor Street, Queen, Christie/Strachan and Coxwell. https://ow.ly/7CW450ZOrXG

bryson reposted

Replying to @shibbi.me

I see where you are coming from! That feels reasonable, but I am still curious of two further questions: whether @mmitchell.bsky.social thinks LLMs can do math and if her readers/sharers of the piece think that LLMs can do math after Colin's evidence

Brian Merchant@bcmerchant.bsky.social

But LLMs are, to this day, still well enough understood as stochastic parrots. It's an accurate description of the technology. Many AI models involve more than LLMs, as fellow co-author Margaret Mitchell pointed out, but it's *still* a useful critical framework for understanding AI's limits.

impeach him on this

Roger Parloff@rparloff.bsky.social

Hearn's case presents "exceptional" circumstances. He "faces the unique possibility that the govt will revive this case—one that it has already announced to be meritless—based on political pressure brought by the President of the United States on his subordinate officers."

bryson reposted

Really useful chart from @graphicmatt.com newsletter about where Mayor Chow's administration has spent more money beyond inflation and population growth, and for what. Good basis on which to choose whether one supports her priorities. Note here were also spending reductions elsewhere.

artificial intelligence has always been an imperfect was to describe the technology maybe we should change it *the monkey's paw curls*

well that's a bad sign

SE Gyges@segyges.bsky.social

I apologize for my earlier article, which did not undergo proper legal review, and will not be able to be present on social media for a while.

this is the best use of LLMs today: finding primary data sources hosted online that a human never would/could

Natalie@nataliethenerd.com

there are so many exaggerated headlines over open ai "hacking" medicare in australia, but to me it sounds like it just found a deep web page and we are the idiots?

wtf

Emma Yeomans@yeomans.bsky.social

Worrying scenes in the Channel today, for all interested in asylum policy and maritime tactics. A small boat from Normandy spent more than 30 hours at sea. Those on board now on their 2nd rescue vessel as Border Force forced to play cat and mouse with far-right activists who have their own boat.

Any 'pacing' regs should mean that all the model labs are mandated to spend the next year building the best tool-use 9B-sized model we can run on-device