My idiotic not-even-briefly-thought-through prediction:
LLM systems can’t do anything reliably, but they can do it cheaply, so ‘decision makers’ contract out LLM based systems to handle customer support/service and slap a disclaimer on it. You can pay for premium support by a person. Your PA still has to wait on hold to talk to this person. If you can’t afford a PA, you can’t afford the premium service.
It’s so much cheaper than having real people, not because it’s cheap, but because now you have infinitely scalable team size with no downtime, no managers, no benefits, or other ancillary costs. Quality goes down. Unit economics improve.
Companies that don’t jump on the LLM train will be more expensive because they have to have real people, which are better on almost every metric, but expensive to maintainne. No one will be able to afford the premium service of having someone who actually has the ability to do anything, so they will lose customer base, so they’ll get more expensive, so they’ll lose customer base, so they… Repeat until they are bought out by their competitors or are a luxury brand charging fifty billionaires whatever they feel like because the check is blank.
The market polarizes. Everything is either barely affordable crap or expensive luxury service. No one is happy, but ‘profit’ is maximised. The American health insurance company becomes the model for everything.
It can’t even do it cheaply. The current API prices are far too low for inference providers to survive long term, because not only are they currently deep in the red on inference itself, the hardware this stuff runs on has a very limited lifetime of a few years, means you have to replace the most expensive parts the most (pretty much different than most things - normally the really cheap stuff has to be replaced regularily, which makes a lot more sense). These costs do not even include the massive bill for training new models, which they all have to do to make sure to not fall behind.
See the first point, in combination that many areas are not tolerant to high error rates - noone hires a lawyer that makes grave mistakes on the regular, and i refuse to see a doctor that messes up 20% of their diagnoses.
Anyone not jumping on the train will be safe when the inevitable moment comes where inference providers start raising prices massively to get into the black, everyone else will be shafted because of the loss of institutional knowledge due to laying off their human staff.
Because the tech is basically ready baked at this point and it only needs training to get up to date on recent events, I suspect the cheaper, possibly open models will be used for a lot of BS. You don’t need the latest vajillion dollar model to run customers through a loop of ‘How can I help you today? … Could you please tell me your account number? … Please wait while this input is processed by an open-source LLM on a raspberry pi CPU. … Can I interest you in signing up for our premium service to receive human support? … That’s okay. I’ll ask again later. … Could you please confirm your account details? … I’m sorry. I’m not authorized to solve your problem. Please visit the nearest branch office of our company, located conveniently one mile from the geomagnetic south pole and our one free tier customer service penguin will be happy to hear your complaint. Is there anything else I can help you with today?’
Doctors and lawyers are professionals with niche skills in reasoning and domain knowledge, not replaceable with an LLM. It takes no knowledge to give someone the runaround. If your goal is the pretense of customer service without actual service, you can run shitty models on shitty hardware to make it slow and even more unreliable, and just slap a disclaimer up front saying your company can’t be held responsible for what the LLM spits out. It won’t matter if it sucks when it cuts your payroll expenses and cranks your shareprice so hard that every competitor follows suit to get their bonuses up, or frees up so much cash you just buy out the competition given anti-trust laws are just ignored these days.
Never doubt the ability of a business schooler to make the world a worse place.
I can only speak to my field, which is legal. Incomprehensible but prolific drivel has always been something the court system has struggled with getting through. That same drivel is now pretty good at pleading generalities but it’s awful at detailing facts.
You can get past the first hurdle in the legal world and unlock a massive and expensive cascade which AI can also effectively navigate for you.
In essence, AI can, very effectively, take an incoherent unreasonable and often mentally unwell litigant and magnify the expense of getting rid of their frivolous claim by over 100.
It’s almost assuredly going to 1. Chock the system. 2. Force massive reallocation of resources in the judiciary 3. Thereby vastly increasing not decreasing legal work. And 4. End up with valid claims being killed in a massive change in judicial attitude toward “clear enough” pleadings as they struggle with an AI influx.
If only. If you google an AI specifically for litigation you’ll get websites selling weekly access for just shy of $500.
I often get very large pleading dumps immediately prior to a hearing instead of at a more logical time and I think it’s the pro se scrounging the money together for another fix of an expensive AI website.
I don’t know. I mean it’s probably a very bad model and these people are just being scammed for a gpt 3.0. I can’t say I like them but I still abhor a scam.
Plaintiff - Submits AI fueled claim.
Judiciary - Uses AI to parse claim.
Outcome: more people attempt claims because previously they were incapable of a cohesive claim, but overall time dealing with individual claims goes down. Probably not in the short term though.
I sincerely doubt it. Judges want to do justice. I think most people would call having a robot judge your pleadings and dismiss merit less claims as dystopian.
Even if it was just “parsing” that’s a layer between you and the judge. Probably also dystopian.
My idiotic not-even-briefly-thought-through prediction:
Business majors will find a way to make it worse.
Because the tech is basically ready baked at this point and it only needs training to get up to date on recent events, I suspect the cheaper, possibly open models will be used for a lot of BS. You don’t need the latest vajillion dollar model to run customers through a loop of ‘How can I help you today? … Could you please tell me your account number? … Please wait while this input is processed by an open-source LLM on a raspberry pi CPU. … Can I interest you in signing up for our premium service to receive human support? … That’s okay. I’ll ask again later. … Could you please confirm your account details? … I’m sorry. I’m not authorized to solve your problem. Please visit the nearest branch office of our company, located conveniently one mile from the geomagnetic south pole and our one free tier customer service penguin will be happy to hear your complaint. Is there anything else I can help you with today?’
Doctors and lawyers are professionals with niche skills in reasoning and domain knowledge, not replaceable with an LLM. It takes no knowledge to give someone the runaround. If your goal is the pretense of customer service without actual service, you can run shitty models on shitty hardware to make it slow and even more unreliable, and just slap a disclaimer up front saying your company can’t be held responsible for what the LLM spits out. It won’t matter if it sucks when it cuts your payroll expenses and cranks your shareprice so hard that every competitor follows suit to get their bonuses up, or frees up so much cash you just buy out the competition given anti-trust laws are just ignored these days.
Never doubt the ability of a business schooler to make the world a worse place.
I can only speak to my field, which is legal. Incomprehensible but prolific drivel has always been something the court system has struggled with getting through. That same drivel is now pretty good at pleading generalities but it’s awful at detailing facts.
You can get past the first hurdle in the legal world and unlock a massive and expensive cascade which AI can also effectively navigate for you.
In essence, AI can, very effectively, take an incoherent unreasonable and often mentally unwell litigant and magnify the expense of getting rid of their frivolous claim by over 100.
It’s almost assuredly going to 1. Chock the system. 2. Force massive reallocation of resources in the judiciary 3. Thereby vastly increasing not decreasing legal work. And 4. End up with valid claims being killed in a massive change in judicial attitude toward “clear enough” pleadings as they struggle with an AI influx.
Ugh. I hadn’t even thought of AI slop litigation. These companies are giving a firehose to the mentally ill as a monthly subscription.
If only. If you google an AI specifically for litigation you’ll get websites selling weekly access for just shy of $500.
I often get very large pleading dumps immediately prior to a hearing instead of at a more logical time and I think it’s the pro se scrounging the money together for another fix of an expensive AI website.
At least that’s prohibitively expensive for a lot of people. It’d be so much worse if it was affordable by more people.
I don’t know. I mean it’s probably a very bad model and these people are just being scammed for a gpt 3.0. I can’t say I like them but I still abhor a scam.
Prediction:
Plaintiff - Submits AI fueled claim.
Judiciary - Uses AI to parse claim.
Outcome: more people attempt claims because previously they were incapable of a cohesive claim, but overall time dealing with individual claims goes down. Probably not in the short term though.
I sincerely doubt it. Judges want to do justice. I think most people would call having a robot judge your pleadings and dismiss merit less claims as dystopian.
Even if it was just “parsing” that’s a layer between you and the judge. Probably also dystopian.