A Letter From The Appraisal Institute

[TerryRohrer]

AppraisalPort is owned by CoreLogic.

AI Ready (the data standard underpinning Appraisal Port) was developed by the Appraisal Institute and FNC, Inc. FNC, Inc. was later acquired by CoreLogic.

 

[alebrewer]

AI Ready (the data standard underpinning Appraisal Port) was developed by the Appraisal Institute and FNC, Inc. FNC, Inc. was later acquired by CoreLogic.

ahhh. yes - that POS was originally developed by AI.

 

[Surf Cat]

Have you ever had an objective review of one of your reports with the intent to identify any deficiencies that, when addressed, will improve all of your reports going forward?

Many times.

It is likely that 99% of practicing appraisers could benefit from something like this. Of course, I would expect 99% of those to react as if someone was criticizing them personally and completely miss the benefit or the spirit in which it is offered.

When you put it in that context, I agree with you.

My comment was directed at GSE work through AMCs. Having been an AMC appraiser, I'm very aware that they don't want the best, most experienced and competent appraisers.

I can see this benefit being very useful for court testimony and complex private work outside of the GSE Realm.

 

[RCA]

The benefit, known as Advisory Guidance for Designated Members, will provide a unique educational opportunity to enhance future appraisal reports!
You may submit an eligible appraisal report for independent advisory guidance from a fellow Designated member, serving as an Advisory Guidance Counselor, who has knowledge and background in evaluating appraisal reports.

What does this have to do with GROK and AI?

I know and have worked with some major reviewers - and they have always said that reviewing takes more time than doing the original report!!!. So, do you REALLY think they are going to REALLY review reports for free? Really?

No, more likely, they are looking for reports from SRAs to train LLM systems on. And to cut down time, they will most likely run them through some LLM to cut down the time for review. Most probably - I would guess.

 

[RCA]

Why would an Appraisal Institute member with an SRA designation need advisory advice from another member for an appraisal?

I mean..... an SRA has to pass advanced courses and case studies to become an SRA. I could see this benefit if the SRA was moving up to CG licensure.

Everyone and anyone who has worked for AMCs knows that the AMCs don't hold an SRA in high regard. They want the newbies to control. Just look at some of the recent thread posts here..... the most recent one being withdrawing from an assignment because there were no comps in close proximity to a dump site. A basic assignment deemed too complex because the newbie didn't know how to do it.

You can bounce a complex situation off other members here on the AF for free.

As already posted, they are most likely looking for relatively good reports to train their LLM on. Now, of course, they already have some SRA reports likely in their archives, but there may be legal issues with accessing those reports. And as I said, it is not likely that anyone is going to be doing this work, without getting something in return. -> Common Sense.

 

[J Grant]

I know and have worked with some major reviewers - and they have always said that reviewing takes more time than doing the original report!!!. So, do you REALLY think they are going to REALLY review reports for free? Really?

No, more likely, they are looking for reports from SRAs to train LLM systems on. And to cut down time, they will most likely run them through some LLM to cut down the time for review. Most probably - I would guess.

You invented that reason why they are reviewing reports for free - they might have offered it to you or others who perhaps got pushback on a report, however, idk the reason why -they must be aware taht they are putting themselves out of a job by training a system, plus imo AI ( artificial intelligence ) is over rated and will likely always need intense human intervention -to function in a way that benefits us.

I am already experiencing problems with AI in spell check and grammar, as it replaces meaningful words with idiot words - it works very well for some professions and some functions within a profession but as a holistic whole, replacing a human, - I don't see it , at least not for a decade or more if ever.

 

[RCA]

Have you ever had an objective review of one of your reports with the intent to identify any deficiencies that, when addressed, will improve all of your reports going forward? In the early days of our Board, the State employed a licensed CG and every year they randomly selected appraisers who had to submit a report upon request that was then reviewed by the state. Never comfortable, but the few I sent in were dinged for minutia and knowing that the person who would review my reports when complaints were filed was not finding glaring errors was useful information. Of course, when Board members needed to lower the bar so they could get over it, that ended.

Since then, I have hired appraisers who I was certain would be objective and tell me what I needed to hear, rather than what I wanted to hear, to review my reports to ensure that I wasn't missing anything glaring, no matter how small, that could be a liability in court or before the Board. I'm not sure why this is being criticized. It is likely that 99% of practicing appraisers could benefit from something like this. Of course, I would expect 99% of those to react as if someone was criticizing them personally and completely miss the benefit or the spirit in which it is offered.

Except, keep in the back of you mind, that these reviewers will most likely be passing your report through and LLM ---- and these LLMs are progressing very rapidly --- you would be surprised how easily they can find errors in PDFs: Typos, the minor errors, logic errors, and even criticize the strategy and approach used. But hey - if you can ask Grok-3 or some other LLM the right questions (very important!) you might be surprised what it comes back with. And, I would, as a warning, recommend you run any report through a good LLM before sending it off to the AI, which will likely do the same--and use it for training a special LLM for reviews.

So, we have Grok-3 and they are saying they are about to come out with a Grok-4 and Grok-5 in the near future.

Claud has just upgraded from Claude 3.5 to 3.7 (which many thought would be 4.0 --- but they are already working on a more advanced version for 4.0).

They may be using the ChatGPT-01 Pro - which is $200/month!!

They may be training their own Deepseek setup. But, who knows.

Look at the good and bad. You can run your reports through an LLM to do your own review and make your report better. Of course the reviewer can will probably have a better LLM to review your report. But that is where this is going.

It is, at this point, just common sense dictating the path to the future.

 

[RCA]

Many times.

When you put it in that context, I agree with you.

My comment was directed at GSE work through AMCs. Having been an AMC appraiser, I'm very aware that they don't want the best, most experienced and competent appraisers.

I can see this benefit being very useful for court testimony and complex private work outside of the GSE Realm.

Everybody has access to LLMs. - Including attorneys.

 

[Non Sequitur]

The institute owned appraisal port. Do they still own it. At one time there was talk of an institute AMC.

There was a licensing agreement between FNC and the institute for the old AIRD. I can't remember the full story about institute's AMC but IIRC it was spun off but not to FNC.

 

[TerryRohrer]

Except, keep in the back of you mind, that these reviewers will most likely be passing your report through and LLM ---- and these LLMs are progressing very rapidly --- you would be surprised how easily they can find errors in PDFs: Typos, the minor errors, logic errors, and even criticize the strategy and approach used. But hey - if you can ask Grok-3 or some other LLM the right questions (very important!) you might be surprised what it comes back with. And, I would, as a warning, recommend you run any report through a good LLM before sending it off to the AI, which will likely do the same--and use it for training a special LLM for reviews.

So, we have Grok-3 and they are saying they are about to come out with a Grok-4 and Grok-5 in the near future.

Claud has just upgraded from Claude 3.5 to 3.7 (which many thought would be 4.0 --- but they are already working on a more advanced version for 4.0).

They may be using the ChatGPT-01 Pro - which is $200/month!!

They may be training their own Deepseek setup. But, who knows.

Look at the good and bad. You can run your reports through an LLM to do your own review and make your report better. Of course the reviewer can will probably have a better LLM to review your report. But that is where this is going.

It is, at this point, just common sense dictating the path to the future.

I doubt there is sufficient volume involved in this endeavor to result in much training, but I guess they all start somewhere.