Hey, @PabloTorre I see you got a clear schedule now. Lemme know your thoughts on guys by the name of Nico Harrison and Robert Pelinka.
Iโll pay whatever is needed and I love your work.
@Jason Working on something aimed at bridging the information gap between homeowners and contractors.
It helps homeowners better understand bids, contracts, and what questions they should be asking before they sign anything.
I think renovations are another part of this. There are construction signs all over my neighborhood in Dallas for people adding on to homes they already own instead of moving.
When you already like your house and neighborhood, that can make a lot of sense, especially when adding square footage can still be cheaper on a $/sf basis than buying a larger home.
Really enjoying your data on this, Jason.
I think a big part of this is just how much housing Texas has built.
In places like DFW, Austin, San Antonio and Sherman, existing-home sellers are competing with a lot more supply, including new construction with builder incentives.
Mortgage rates and local job markets obviously matter too, but Iโd assume all that building is playing a pretty big role. More supply should mean more competition and more pressure on prices.
Iโd assume regulators can still put requirements on the providers distributing the models, but once someone is self-hosting and modifying one, enforcement gets a lot murkier.
Thatโs really what Iโm curious about - how far can these rules realistically reach once open weights are out in the wild?
@simongarlick@Pv If this is coming from the EU, why are the labs applying it globally instead of just there?
And is the main point really just being able to tell later whether something was AI-generated?
@KUBLAIxxx@evandotsh@simongarlick@Pv Thatโs what Iโm curious about too. A lot of companies already handle regulations differently by market, so why are the labs applying this globally instead of just in the EU?
@NoahRettberg@Polymarket I get that itโs required - Iโm just trying to understand the point of the requirement. Is it mainly about being able to identify AI-generated content later, or is it supposed to prevent something more specific?
@brainzcode@Pv If this is really being driven by regulation, do we think open source stays untouched? Iโd imagine regulators eventually try to apply some version of the same requirements there too.
@jamesExplainsIT@Pv@AnthropicAI@OpenAI If that happens, does open source become the obvious alternative for people who donโt want watermarking? And can regulators realistically reach locally hosted/open-weight models the same way they can the big labs?
@CoreyNoles Completely agree. AI should be treated like any other tool - the output still has to stand on its own. If the work is good, useful, and accurate, I care a lot less about whether it was done with AI, Excel, or a pen and paper.
@Pv Is this mainly a transparency thing, or could it eventually affect ownership/use of the output too?
I get the concern if watermarking limits the model, but if models keep improving Iโd think that tradeoff matters less over time.
@KingFazir Iโm seeing this may be driven by new EU transparency rules. If thatโs true, do all the frontier labs have to implement something similar, or can they comply in different ways?