as he said at the time -- the NCSA environment as incubator for Netscape builds on the "tradition" of university-supported student innovation that "has been a long-term positive aspect, and indeed a driving force, of the American economy" groups.google.com/g/comp.infos...

also they had REU grants specifically right around the time he worked there as an undergraduate programmer (reported as 1991-1994). i dunno if his personal salary/expenses were directly paid by an NSF grant, but the existence of the project & student staff was deeply tied to NSF funding streams

re @adamgurri.liberalcurrents.com on NSF support for the web www.liberalcurrents.com/marc-andrees... --

the NSF specifically gave $3.7M just for Mosaic
www.nsf.gov/awardsearch/...

and $4.6M for a related effort
www.nsf.gov/awardsearch/...

yeah! though i'm iffy on their concept of "information". their example of socially transmittable body attributes - for sure part of culture, i dunno information.

ah i mean all the uses for stuff that's not (academic) research. like Fig 3 here (2024?). there's.. lots, i guess. arxiv.org/pdf/2309.11998

This statement reminds me of each after all the Meta/FB scandals - when they harm users, or even cause genocide, they just blame their internal automated systems for being imperfect.

from
www.nytimes.com/2025/08/26/t...

bundestag elections & pca github.com/brendano/bun...

i still can't believe how well latent space models work for political ideology. even just on 12 datapoints here.

And NSF's mission to support diversity is in CURRENT law!

42 U.S. Code § 1885 and its subsections (a), (b), and (c):
www.law.cornell.edu/uscode/text/...

Congressional intent is explicitly for what's now called DEI.

in springfield, MA

@jeffstein.bsky.social

Suresh Naidu at UMass (Bowles Lecture).

connecting LLMs to socialism (!)

if this hasn't happened yet, we'll have tension b/w wanting models to represent globally diverse perspectives, versus ethical values of the model authors or their societies. e.g. i like this chatgpt response but it's not reflective of global culture:

www.pewresearch.org/global/2020/...

cloud forecast looking great

Good example of how industry AI threatens the credibility of corpus-based social science.

MICASE was clearly intended for pedagogical/ling research (image from linked doc), and sounds nice for that. We need non-commercial restrictions to safely allow this type of data collection & research

Tomorrow at #CRAC2023, Ankita is presenting ezCoref, our open-source coreference annotation tool, and a re-annotation study!
aclanthology.org/2023.finding...

#CRAC2023
@emnlpmeeting
Dec 6, 2:50PM

twitter thread:
twitter.com/Anki98765/st...