Daniel Milo @DMilo75

Finance PhD student at NYU Stern AI, Monetary Policy, Real Estate dan-milo.com Joined June 2022

Tweets

45
Followers

264
Following

1K
Likes

299

Daniel Milo @DMilo75

3 days ago

@toddsaunders They’re still helpful to guide interaction with certain external tools. Multi factor authentication seems to trip it up without a skill explaining how it works for that tool.

0 0 0 138 0

View Details

Daniel Milo @DMilo75

3 weeks ago

@alz_zyd_ It’s also quite helpful to cut through pre-AI slop on the internet, where stretching out exposition and filling it with ads was the norm.

0 0 1 201 0

View Details

Daniel Milo @DMilo75

3 weeks ago

@soumitrashukla9 How many workflows do you have running at once? I still find the human guidance/context switching costs too high to go beyond three or four at once.

0 0 0 84 0

View Details

@alz_zyd_ And yet markets and the Fed spend considerable effort on GDP. One reason: economic measurement is in general quite crude (not representative, measurement error). GDP is one of the few decent signals. With AI, we might get to focus on other things, from better measurement.

0 0 0 48 0

View Details

Daniel Milo @DMilo75

a month ago

@LinkofSunshine This currently happens in California with property taxes

0 0 1 737 0

View Details

Daniel Milo @DMilo75

a month ago

@alz_zyd_ I figured a running list of assumptions, version history, parsed literature, ability to run simulations, etc would improve things.

0 0 0 39 0

View Details

Daniel Milo @DMilo75

a month ago

@alz_zyd_ Why do you think it’s not more helpful in directly developing theory?

1 0 1 397 0

View Details

Daniel Milo @DMilo75

a month ago

@JeremyWS The people you talked to that grumbled the Fed communicated too much, what reasons did they give?

1 0 0 127 0

View Details

Daniel Milo @DMilo75

a month ago

@zaeemal_ It’s a consequence of how quickly agents have improved. Once the dust settles we’ll have new ways of doing research that keep us more cleanly in the loop.

1 0 1 33 0

View Details

Daniel Milo @DMilo75

a month ago

@bswud In the US we started with far simpler regulation, perhaps explaining our early growth, but ended up in the same spot. More regulation seems the natural progression, it's the easier way to satisfy people in the near-term at the expense of long term progress.

0 0 0 8 0

View Details

Daniel Milo @DMilo75

a month ago

@mycelias On the flip side, governments may use AI themselves to write longer and more complex regulations/processes. Still unclear where the equilibrium will be.

1 0 1 50 0

View Details

Daniel Milo @DMilo75

a month ago

@achenfinance Private data (especially in econ/fin) papers will also make this challenging

0 0 0 199 0

View Details

Daniel Milo @DMilo75

a month ago

@iammattduff @barrowjoseph @DenisPeskoff DM’d

0 0 1 27 0

View Details

Daniel Milo @DMilo75

a month ago

@mycelias @HeySilasReed @barrowjoseph @DenisPeskoff @chiefofscouts @engagicc How far back in time do you go? Also, can you track variance requests/decisions?

1 0 1 48 0

View Details

Daniel Milo @DMilo75

a month ago

@HeySilasReed @barrowjoseph @DenisPeskoff @chiefofscouts The underlying text is public and can be hosted, but it's more challenging to collect and parse PDFs (which nobody continually maintains AFAIK) vs use the structure aggregators (i.e. municode) provide. We're working on the FOIA route, but its costly too.

1 0 2 61 0

View Details

Daniel Milo @DMilo75

a month ago

@TheStalwart I was excited about LLM tournament scoring at first, but it seems to produce very similar results to asking outright for a score (say 1-10), and costs more/is more complex to understand.

0 0 0 59 0

View Details

Daniel Milo @DMilo75

a month ago

@readforyourself @barrowjoseph @DenisPeskoff @DDemocracy_ca @CalPolyIATPP During work on thezoningproject.org, we found some states (especially in the southeast) to charge money for document requests and/or only work with in-state residents. Another hurdle is potential legal ownership from aggregators like Municode.com