Anyone Using Aivre?

[cashcatz]

The last time I tried it for mapping and sales data analytics AIvre required latitude and longitude to be imported from the MLS which Oregon RMLS does not have.

How long ago was that? That's literally as simple as running the normalized address across a geocoder - the US Census Bureau even offers one for free via API.

 

[OSU Beavers]

It was choking hard on Oregon RMLS export two months ago. I lost interest after hearing the $3000 a year price tag for AI that hallucinates.

 

[RCA]

Reports I have seen from appraisers trying to use AIVRE are that the AI doesn't work well, ranging from the automated data entry requiring extensive review and revision and the trend analysis not working at all. I haven't heard where the development of their mobile app is lately. Last I saw, it didn't exist.

I definitely believe that. They don't provide any substance in their sales documentation. They are estimating what they will have by 2027. What they likely have is Random Forests level regression. That is Black Box stuff. You can't use that for standard appraisals.

No, there is really no way a national or international company based in New York City can provide genuine appraisal services for the entire country. Maybe they are waiting for me to come out with the next phase of my software. But the more advanced software like MARS (or "R/earth") is even more time-consuming than the traditional approach, at least until you have entire counties locked and loaded (fully understand and modeled) and in a constant state of maintenance/updating. That is, even if you have the models for all property types, they have to be continually updated.

They want $3,500/year for the Ultra? That is about $300/month. That is exorbitant. In any case, they can't deliver, at least until they can start providing the robots to do the leg work.

They don't know what they are doing. They are being premptive in what they think they can do. But someday they will find out it is not that easy.

Now, they are highly qualified educationally, and they may have a few MAIs in their midst, but I don't see any indication that, at this time, they know what they are doing with respect to core appraisal, to accurately estimating market value with advanced methods.

 

[TerryRohrer]

I definitely believe that. They don't provide any substance in their sales documentation. They are estimating what they will have by 2027. What they likely have is Random Forests level regression. That is Black Box stuff. You can't use that for standard appraisals.

No, there is really no way a national or international company based in New York City can provide genuine appraisal services for the entire country. Maybe they are waiting for me to come out with the next phase of my software. But the more advanced software like MARS (or "R/earth") is even more time-consuming than the traditional approach, at least until you have entire counties locked and loaded (fully understand and modeled) and in a constant state of maintenance/updating. That is, even if you have the models for all property types, they have to be continually updated.

They want $3,500/year for the Ultra? That is about $300/month. That is exorbitant. In any case, they can't deliver, at least until they can start providing the robots to do the leg work.

They don't know what they are doing. They are being premptive in what they think they can do. But someday they will find out it is not that easy.

Now, they are highly qualified educationally, and they may have a few MAIs in their midst, but I don't see any indication that, at this time, they know what they are doing with respect to core appraisal, to accurately estimating market value with advanced methods.

I think you are giving them more credit for appraisal acumen than they are due. The main man/appraiser started appraising in 2014 and became certified in 2020. I have never seen or heard of them having any other appraisers in the mix. The other co-founder is ProtoCoding. The name strikes me as based on "prototype" which would be to develop a functioning model and sell for a big payday. I suspect that is the intent. Other than the cost and the fact it doesn't work, I am troubled by the fact that there are no details about the "company" behind it all. It is almost certain to fail, so prospective users should think through how they react and survive when, on cold morning, they can't log into their account. Remember ANOW, the last latest and greatest? Other than a few lingering lawsuits, there is little remaining evidence they existed.

 

[Bobby Bucks]

.



They're a Platinum Sponsor at the Tinker Wagon Valuation gathering. We can be comforted in knowing they were vetted by elite industry professionals.

 

[Bobby Bucks]

..........................

 

[TerryRohrer]

Here is the money. Almost three years in existance.

"Draper Goren Blockchain

Draper Goren Blockchain (DGB) is an early-stage hybrid venture studio and venture fund that incubates, accelerates, and invests in startups operating at the intersection of blockchain infrastructure and artificial intelligence. [1, 2]

Leadership Team
The firm was founded in 2023 by a prominent team of Web3 investors and entrepreneur
"

 

[Bobby Bucks]

Never heard of Craftsman, is that building costs? Tell me about then MLS tools, are these your custom spreadsheets?

Craftsman is for building cost. MLS tools include demographics, various market statistics, subdivision statistics, market & zip code sales and trend reports. I prefer depreciated cost, but every few months will use excel. For the truly complex multi million dollar singewides, I refer to the old reliable source “ the list”.

 

[J Grant]

I have this amazing software that is free to use, portable, and available 24/7. It can select comps, analyze data, complete a sketch, and write a narrative. It is called my BRAIN.

 

[RCA]

I think you are giving them more credit for appraisal acumen than they are due. The main man/appraiser started appraising in 2014 and became certified in 2020. I have never seen or heard of them having any other appraisers in the mix. The other co-founder is ProtoCoding. The name strikes me as based on "prototype" which would be to develop a functioning model and sell for a big payday. I suspect that is the intent. Other than the cost and the fact it doesn't work, I am troubled by the fact that there are no details about the "company" behind it all. It is almost certain to fail, so prospective users should think through how they react and survive when, on cold morning, they can't log into their account. Remember ANOW, the last latest and greatest? Other than a few lingering lawsuits, there is little remaining evidence they existed.

There is no "substantive" evidence. Just fluff. We have a lot of that around.