Modern LLMs are incredibly good compression algorithms, which can shed light on why autonomous data science agents don't overfit as much as you might think. arxiv.org/abs/2606.11045
And today, on the centenary of Gaudí's death (10 June 1926), the central tower is to be inaugurated (it was actually topped out on 20 Feb)
For fans of Talkie-1930 and all the methodological questions raised by historical models, here's a new entrant into the field, TypewriterLM, trained up to 1913. Corpus, instruction-tuning datasets, and event dataset are released. arxiv.org/abs/2606.02991
We introduce TypewriterLM, a 7.24B History language model (LM) trained exclusively on English text predating 1913. Developing History LMs requires addressing challenges in data quality and availabilit...
Is something “too dangerous” if it’s easy to slap on guardrails?
Tigers are really dangerous.
Tigers are at the zoo in my town. Don’t go in the tiger enclosure at the zoo. The zoo is pretty good at keeping you out. Under normal circumstances, tigers aren’t really something to be alarmed about.
Today La Sagrada Família is having the first part of the crowning cross of the Jesus Tower installed. This will finally take it past Ulm Minster's 161.5 metres (begun 1392, completed with octagon and spire 1885-90) as the tallest church in the world!
Hey everyone: this is important: please take a few minutes to post a comment objecting to this brazen attempt by #Trump administration to politicize ALL science:
www.regulations.gov/document/OMB...
I feel I'm being lowkey trolled when a company puts out a chatbot that is known for hallucinations/confabulations and calls it .... Fable
Here is Sen. Warren in Time Mag correlating mass layoffs to the current US tax code.
“Right now, companies pay payroll taxes for their workers but get tax breaks for investing in technology—effectively, a tax penalty for hiring human beings and a tax break for buying equipment.”
“Reacting to the GAAIA (Great American Artificial Intelligence Act), Samir Jain, vice president of policy for CDT, told Inside AI Policy the ambiguity around the preemption provision creates an opportunity for entities to game the system….” insideaipolicy.com/s...
