alexcannan
Re:OpenAI is Visa – Buttering up the government to retain a monopoly
on Thursday December 26, 2024 @07:44PM (#42519048)
Nothing in this article is anti-tech. Anti-monopoly, yes.
Re:Spotify demonetizes all tracks under 1k streams
on Monday April 08, 2024 @09:24AM (#39969482)
Computers are very good at doing these calculations
Re:Yale will again require standardized test scores for admission
on Thursday February 22, 2024 @10:36AM (#39468632)
Surrounding yourself with the smartest people possible, during the most formative years of your life, is always a good idea. Get into the best school you can
Re:American Universities Are Post-Truth
on Saturday January 13, 2024 @09:45AM (#38980363)
I think any kind of decision by a university that can be framed as "social justice" will be, and then be amplified by the media without end, no matter how small. Our institutions are so much more than a handful of decisions at the executive level (which are probably aligned with popular support of people who actually attend the university, go figure).
Re:Turo Wants Customer to Pay $35,000 After Vandals Smash His Rented C...
on Wednesday October 25, 2023 @04:22AM (#38010524)
I had a similar experience with Turo. The owner instructed us to leave the car unlocked in some parking lot. The following day they sent pictures of pot on the floor of the car; they charged me an extra $100 and handed me a lifetime ban. I was really enjoying using the app too...
Re:Single Digit Tokenization improves LLM Math abilities by up to 70x
on Thursday October 19, 2023 @01:33AM (#37938870)
I'm personally excited about LLMs becoming more number-fluent for regression tasks, e.g. read this paper and give it a grade 0-100
Re:We Don’t Need a New Twitter
on Thursday August 17, 2023 @10:40AM (#37162391)
Zero mention of the increasingly popular federated alternatives just seems like bad, corpocentric journalism
Re:Show HN: Scribbler – Podcast Summaries Using GPT
on Wednesday June 14, 2023 @10:26AM (#36326524)
Very cool! I'm curious--I'd imagine that some long tail podcasts have transcripts that are too long to fit within a standard context window. Do you have some strategy for handling these?