As an example, creating recipes with Claude Opus based on flavor profiles and preferences feels magical, right up until the point at which it can't accurately convert between tablespoons and teaspoons. It's like the point in the movie where a character is acting nearly right but something is a bit off and then it turns out they're a zombie and going to try to eat your brain. This note taking example feels similar. It nearly works in some pretty impressive ways and then fails at the important details in a way that something able to do the things AI can allegedly do really shouldn't.
It's these failures that make me more and more convinced that while current generation AI can do some pretty cool things if you manage it right, we're not actually on the right track to achieve real intelligence. The persistence of these incredibly basic failure modes even as models advance makes it fairly obvious that continued advancement isn't going to actually address those problems.
It often feels like the AI industry is continually glossing over the fact that capability and reliability are fundamentally different qualities. We tend to use "accurate" and "reliable" interchangeably, but they describe different things. A model can ace a benchmark (capability/accuracy) and still be a liability in production (reliability).
Just look at recent reactions to yet another release from METR showing improved capabilities. But the less talked about part is how their measure is for a 50% success rate (and the even lesser talked about secondary measure they have at 80% success rate has a drastically lower time-horizon for tasks). https://metr.org/
I implement AI systems for enterprises and I don't know any that would ever be okay with 80% reliability (let alone 50%).
When I see how LLMs are capable of essentially prompt and context engineering for themselves, it makes me think they won't need human guidance forever.
When it comes to simple fact-based tasks that have a concrete methodology, it is no surprise to me that LLMs aren't the right tool, and I believe it's a failure of the harness to not recognize those types of tasks and handle them with a more concretely functioning tool instead of relying on statistical probabilities in the LLM "brain" to spit out the correct number to a math problem.
In the same sense that LLMs can use "skills" when necessary, it should have tools or possibly even specialized "brains" for it to pass of certain types of tasks to.
I'm starting to feel that our first form of AGI is not going to be a single brain but an elaborate system of harnesses, multiple LLM models, skills, domain and task specialized subsystems it passes tasks off to, etc. Whether we get there with current LLM technology before some other evolution in AI is the question, to me.
The main problem currently with LLM text is not that they create incoherent sentences, it's that what they purport to be statements of fact or general consensus often times aren't, because they are bullshit machines that become better and more accurate bullshitters the more context-accurate data they are fed. AI videos may still have issues with "looking plausible" whereas LLM text currently has less issues with "sounding plausible" and more issues with "being correct" with respect to reality. Which they have no direct connection to.
No one is penalizing an AI video generator for creating a scene that never happened in real life.
(If it's saying 3.14tsp or 2tsp then I have no idea)
I got some inspiration from it but it misinterpreted very basic stuff. might be a skill issue on my side, I do not know.
Real intelligence means you have to say "I don't know" when you don't know, or ask for help, or even just saying you refuse to help with the subtext being you don't want to appear stupid.
The models could ostensibly do this when it has low confidence in it's own results but they don't. What I don't know if it's because it would be very computationally difficult or it would harm the reputation of the companies charging a good sum to use them.
I think they're getting better at it, but it's likely just the number of parameters getting bigger and bigger in the SOTA models more than anything.
They don't like hearing "I don't know"
I have met many supposedly intelligent, certainly high status, humans who don't appear to be able to do that either.
I have more confidence we can train AIs to do it, honestly.
"Give me your answer and rate each part of it for certainty by percentage" or similar.
So instead of an LLM trying to answer a math or reason question by finding a statistical match with other similar groups of words it found on 4chan and the all in podcast and a terrible recipe for soup written by a terrible cook, it can use a calculator when it needs a calculator answer.
It just feels like for some reason this is all being relearned with LLMs. I guess shortcuts have always been tempting. And the idea of a "digital panacea" is too hard to resist.
You ask an LLM "What's wrong with your answer?" and you get pretty good results.
In other cases, I have seen it miss the mark when the discussion is not very linear. For example, if I am going back and forth with the SOC team about their response to a recent alert/incident. It'll get the gist of it right, but if you're relying on it for accuracy, holy hell does it miss the mark.
I can see the LLM take great notes for that initial nurse visit when you're at the hospital: summarize your main issue, weight, height, recent changes, etc. I would not trust it when it comes to a detailed and technical back-and-forth with the doctor. I would think for compliance reasons hospitals would not want to alter the records and only go by transcripts, but what do I know...
Diagnosed with Runner's Knee.
AI summary said I was diagnosed with osteoporosis, and had hip pain and walking difficulty, though literally none of that was ever said or implied.
CHECK YOUR TRANSCRIPTS. Always, but especially with LLM transcribers, which fairly frequently include common symptoms which don't exist, or claim a diagnosis which is common and fits a few details but not others. Get them fixed, it can very strongly affect your care and costs later if it's wrong.
Anecdotally, I'd say that outside of a couple very simple and very common things, about 50% of the "AI" summaries I've had have been wrong somewhere. Usually claiming I have symptoms that don't exist, occasionally much more serious and major fabrications like this time.
LLMs are NOT normal speech to text software, and they shouldn't be treated like one. They'll often insert entire sentences that never occurred. In some contexts that might be fine, but definitely not in medical records.
Someone else who couldn't attend the meeting later read that summary and it created a major argument because the topic had been a sore subject for this person due to an ongoing debate at the company. Everyone who attended the meeting confirmed it was an error, but the coincidental timing made it hard for him to accept, because the LLMs summary presented things in a way that validated this person's concerns that had been previously minimized by some folks on that meeting.
The drama got heated to the point where management produced a policy about not trusting generative output without independent verification. Seems at least it was a lesson learned.
There are some good uses for AI, but I'm not convinced that this (or many other cases where accuracy matters) is one of them.
- some human checking all the notes by listening to the entire meeting recording (takes a lot of time and man-hours)
- attendees checking notes from memory (prone to error unless they take notes)
- attendees cross checking with their own notes (defies the point of having the AI note taker)
The reality is that AI usage is not acceptable in any form in any context where accuracy is critical, but good luck getting anyone to acknowledge that.
“Notice: Any comments made by <name> or on behalf of <organization> that are interpreted by AI in this meeting, may not be accurate.”
I do this in every meeting.
Freeing up doctor time, for example: lots of patient visits are messy, the patient is scattered, has multiple issues, and the doctor has tight timelines and regulatory challenges to convey to the patient impacting their care… this is architected for everyone to lose, IMO, even with a perfect transcript. And LLMs can’t be perfect, they auto complete.
I picture patients interacting with an intake AI who can listen to hours of demented rambling, or a patient mid anxiety attack, and provide a caregiver-certified summary of needs, with relevant screening information laid out for doctor confirmation. At that point, helpful information about drug access or insurance policies can be presented, for doctor confirmation, to a patient who can clarify and refine their understanding of the system without time pressures.
Elevating the quality of dialogue so the doctor is more focused on the patient, and the patients dialog needs don’t overwhelm treatment. A lot of medicine is filling out forms and checklists, I think auto-complete could create efficiencies in how we fulfill that.
She is a great doctor and thankfully does this due diligence. But it gives me the impression this is forced on doctors without even them wanting this.
I would expect an "AI Note Taker" to faithfully transcribe the entire conversation. With the same quality I see in a lot of automated video subtitles.. ie they use the wrong word a lot but it's easy to tell what they mean by context.
Are these tools instead immediately summarising the whole thing, and that summary is the artifact? Because that is a beyond insane way to treat human communication.
> I would expect an "AI Note Taker" to faithfully transcribe the entire conversation. With the same quality I see in a lot of automated video subtitles.. ie they use the wrong word a lot but it's easy to tell what they mean by context.
That's a reasonable expectation, but would not be a safe one. All transcription tools are not made the same. First it depends on what kind of STT/ASR (speech-to-text / automatic speech recognition) model they are using. A lot of tools like to use some flavor of OpenAI's Whisper model. It works well generally but I would never use it in a critical use case like healthcare. Because it can hallucinate. That's specific to its architecture and how it was trained.
There's a fairly large variety of architectures that can be used for STT/ASR. Some of them are designed for "offline" / "batch" / pre-recorded audio. Some are designed for fast real-time streaming transcription.
There are more factors too like training data. And not just demographics of the speakers in the training data but audio environments too. Was the model trained on echo-y doctor offices with two people being recorded from a crappy smartphone mic or desktop mic? (It could've been! But it's an important distinction.)
And there's more factors than that, but you get the picture (e.g. are they trying to "clean up" the transcript afterwards by feeding it to an LLM, are they attempting to pre-process audio before transcription also in an attempt to boost accuracy)
There's a lot of ways to do it, meaning, there's a lot of ways to screw it up.
Obviously this results in hallucinations, mistaken implications, & inaccurately assumed context.
makes me wonder what quality software the ministry would push (probably mostly qualifications like SOC).
This is apparently this list of approved vendors
https://www.supplyontario.ca/vor/software/tender-20123-artif...
If we just postulate that the systems have a high error rate, I wonder why they are being adopted. They seem extremely easy to test, so I don't see why doctors or hospitals or governments should be getting tricked into buying them if they suck.
From the article: "While 30 percent of a platform’s evaluation score depended solely on whether they had a domestic presence in Ontario, the accuracy of medical notes contributed only 4 percent to the total score."
Accuracy wasn't really part of the scoring, Ontario doesn't care about it.
Not mentioned, as far as I can see: the comparative human mistake rate.
Having seen a lot of medical records, 60% sounds about normal lol.
Or do they use traditional voice recognition algorithms to do that part and then just "fix" the result to look plausible? Which with good quality output might not be much, but with bad can be absolutely everything.
If it is later seems to me that issues will absolutely happen.
She called me back later that night and we chatted for bit and then she paused and sort of uncertainly was like “So… was there something you were needing to tell me?” And I was completely baffled and was like “Uhhhh I don’t think so…?”
She then explained the notification she got about my call and apparently the LLM summary of my voicemail converted a message consisting of 75% well-meaning but insignificant interpersonal human filler (like most voicemails) into this stilted, overly formal business-y speak with a somewhat ominous tone. Assigning way too much significance to each of the individual statements in the message about wanting to talk (to say happy Mother’s Day), inquiring about her availability ASAP (to say happy Mother’s Day) etc. Plus grossly exaggerating the information density of the call making it sound like I left this rambling, detailed message about needing to tell her something that was left completely vague, but possibly important and also time critical.
Added up it made her a little worried when she read it and made me a bit pissed that was the end result of my wishing her well. Because apparently everything needs a half baked LLM summary crammed into it now.
ALWAYS check your summaries immediately, and contact your doctor ASAP. They can generally fix it themselves, and it's best done when everyone still has some memory of the event.
I'm puzzled by this as well. Why not just generate a transcript and be done with it? If it's a particularly long transcript that's being referenced repeatedly for whatever reason let the humans manually mark it up with a side by side summary when and where they feel the need. At least my experience is that usually these sort of interactions don't have a lot of extraneous data that can be casually filtered out to begin with. The details tend to matter quite a lot!
The businesses offering these services want to say "we are using AI" to their stake holders and the government committees who approve this shit don't have the skills or knowledge to evaluate the effectiveness in addition to the fact they likely don't even use the tools they have approved for use.
Transcription is both too good, and not good enough. The magic generative content only makes it worse.
Too good: a lot of commercial settings forbid persistent transcription because it makes an easily discoverable record of specific details. Thats a business risk that can be mitigated simply by having participant notes or summaries where the secretary can omit sensitive discussion or present consensus without specifics. And notes/summaries also introduce a interpretive defense with some “strategic ambiguity.”
Not good enough: if you look at STT its still probabilistic. The actual evaluation output will have just much data about alternate words/phrases as the selected choice. That leaves lots of room for creating alternate impressions or representing words that werent actually spoken. The fact that people _think_ a STT transcript is authoritative only makes this worse.
When you add generative inference in top (eg summarization) you exacerbate both problems. I suspect that counsel is more accepting of summaries as its less likely to contain specific discoverable terms, likely to diffuse responsibility and specificity, and your judge/jury will be more amenable to “the ai summary is wrong” than “the transcription selected the wrong vowels.”
> Medication errors were common (nearly 1 of every 5 doses in the typical hospital and skilled nursing facility). The percentage of errors rated potentially harmful was 7%, or more than 40 per day in a typical 300-patient facility. The problem of defective medication administration systems, although varied, is widespread.
https://jamanetwork.com/journals/jamainternalmedicine/fullar...
> In all, 91 unique studies were included. The median error rate (interquartile range) was 19.6% (8.6-28.3%) of total opportunities for error including wrong-time errors and 8.0% (5.1-10.9%) without timing errors, when each dose could be considered only correct or incorrect
(And if you already see 60% error rates in standard, pre-AI note taking, how does that not translate into many deaths and injury? At least one country's health system in the world should have caught that)
Presumably most doctor's visits are a one-problem-one-solution-one-doctor type of thing. Done deal, notes are never read again. So that alone would explain why high rates of errors doesn't result in injuries or death very often.
Any injury or death caused by poor notes would have to occur when mistakes are done if you're followed for a serious chronic condition, or if you're handled by a team where effective communication is required.
Because most of it is just written down and never looked at again until there’s a lawsuit or something.
The management human who offered the bad tool to the other human is responsible.
The robot cannot be responsible in place of us.
Does it?
What do you base this on?
As someone who can both see the amazing things genAI can do, and who sees how utterly flawed most genAI output is, it's not obvious to me.
I'm working with Claude every day, Opus 4.7, and reviewing a steady stream of PRs from coworkers who are all-in, not just using due to corporate mandates like me, and I find an unending stream of stupidity and incomprehension from these bots that just astonishes me.
Claude recently output this to me:
"I've made those changes in three files:
- File 1
- File 2"
That is a vintage hallucination that could've come right out of GPT 2.0.
If I take the Bitter Lesson into account, I would frame this as needing more focus on enabling a general intelligence to use tools more effectively and when appropriate to essentially stop making mistakes.
A basic example being a calculator. The AI needs to recognize its default thinking pattern doesn't work well for math calculations, so it delegates it to the available calculator tool / skill / MCP instead. An LLM should not be relying on LLM prediction to give a mathematical resultant figure, ever. It should come from a deterministic tool. If anything, the LLM may interpret the problem and convert it into starting math figures to use for calculation.
If we can enable AI systems to learn and apply that for themselves, and even develop their own deterministic tooling and sense of what tool to use for what job, that starts to sound promising to me.
Skills feel like a conceptual stepping stone to the next useful abstraction.
I do wonder if people would be pushing AI so hard if their organizations were planning to hold them accountable for mistakes the AI made
I bet if that were the case we'd see a lot slower rollout of AI systems
Which reminds me, I need to figure out how to turn that off.
- Speech recognition and frontier models are continuing to get better at handling these types of conversations across accents, languages and specialties. The trend is obvious and clear here. Compare GPT 4 with Opus 4.7 and there is no contest. I'd even take GPT 5.4 nano over GPT 4 right now. So, yeah, they have been improving and, yeah, they will keep on improving.
- The pipelines these models are being built into are getting much more sophisticated than just 'transcribe with x and have GPT XX clean it up'. The people building these things aren't standing still. Even if they did keep using the same models the pipeline improvements would make things get better over time. Add that in with the model improvements and the gains are even greater.
- The companies doing this work are seeing more and more edge cases. Data matters. More and more practitioners are using these things. That means more to learn from. It also means more stories of things being wrong. If you cut your error rate in half but increase your customer base by 10x then you will be hearing about 5x the problems. We are seeing that right now.
- Providers are starting to adjust to the technology (repeat areas they know may cause trouble, adjust their audio setups, etc etc) Just like any technology both sides shift and it matters. The first users were champions. The second wave were mixed between champions, haters and people that didn't care yet. Now people are really starting to count on this technology. They know it isn't a fad and isn't going away and are actually using it day to day to get their work done. This means they are adjusting to it as needed to get to the next patient/note/etc.
This stuff is just a few years old and the gains are obvious and massive. They aren't going to suddenly stop improving. There is an argument that they will asymptotically approach some level of utility, but we are still gaining quickly right now.
Yes, another layer to cross-check, say, “in kubectl logs I see …” with an actual k8s tool call can help, that is, when the cross-check layer doesn’t lie.
For the time being, IMHO, human validation in key points is the only way to get good results. This is why the tools make experienced people potentially a lot more efficient (they are quick to spot errors/BS) and inexperienced people potentially more dangerous (they’re more prone to trusting the responses, since the tone is usually very professionally sounding).
I'm available for a small fee.
That's because, despite the many claims to the contrary, the models haven't actually gotten any smarter. They are still just token prediction engines at the end of the day, without any understanding of what they are doing. That's why one should not rely on them.
I'm not sure, but it seems to me that if scale or small architectural tweaks were going to solve comprehension, they would have done so by now.
You can.
It just won't do it.
https://chatgpt.com/share/6a06a4c5-d454-83e8-a5b2-c9468f6588...
When you point it out "Oh yes, I did do that which is contrary to the rules, request <whatever>.. Anyway..."
The most common failure I've seen come from tools that pollute their context with crap and the llm will forget stuff or just get confused from all the irrelevant sentences; which if the report is true, is probably what these ai notetakers are guilty of. This problem gets exacerbated if these tools turn on the 1M context window version.
> An alternative way to obtain uncertainty estimates from LLMs is to prompt them directly. One benefit of this approach is that it requires no access to the internals of the model. However, this approach has produced mixed results: LLMs can sometimes verbalize calibrated confidence levels (Lin et al., 2022a; Tian et al., 2023), but can also be highly overconfident (Xiong et al., 2024). Interestingly, Xiong et al. (2024) found that LLMs typically state confidence values in the range of 80-100%, usually in multiples of 5, potentially in imitation of how humans discuss confidence levels. Nevertheless, prompting strategies remain an important tool for uncertainty quantification, along with measures based on the internal state (such as MSP).Yes. This is what medical records are. They've been kept by doctors for a reason.
It's not like the doctor is talking to you about which anime series are the best. You're talking about your health, your body, your disease, your treatment.
It's important to keep track of that.
>Plus what doctor has time to sit down and re-listen to your visit to check to make sure the AI didn't screw up at some point in the future anyway
No doctor.
Which is why it really should be their (or their assistant's) job to record the relevant parts of the conversations.
>At what point does it become a larger waste of time and money to babysit an incompetent AI than just not using one in the first place?
At this point, as the audit shows.
Except the industry (both the AI vendors and healthcare) are going YOLO¹ and relying on AI anyway.
>There are some good uses for AI, but I'm not convinced that this (or many other cases where accuracy matters) is one of them.
This has always been the case, but the marketing now has reached of point of gaslighting in trying to make people collectively forget that or pretend that it's not the case.
Once hard evidence is presented (like in this case), the defense is invariably that it's a temporary quality issue that's going to be resolved as the AI improves Any Day Now™, and that it's wise to live as if it were the case already² (and everyone who disagrees is a fool that Will Be Left Behind™).
The level of fervor in this rhetoric gives me an impression that the flaw is so fundamental that it won't be fixed in any form of AI based on today's technologies, that the AI vendor leadership knows this, and that the entire industry is, at this point, is a grand pump-and-dump scheme.
I hope I'm wrong.
____
¹ See, you only live once. But there are millions of you. So, like, whatever if you don't. Something something economies of scale to them.
² This is called a phantasm.
Not every conversation. Historically, one of the nice things about doctors is that they're the ones filtering what gets included in your medical record. They decide what is medically relevant and what can remain confidential. Doctors understand that not everything discussed needs to be included in your file. Sometimes that really is just small talk, sometimes it's even medical concerns, questions, or requests for advice and still not all of it needs to go into your file and much of it would only clutter it up anyway.
Any system that stores an entire visit as audio or video long into the future (much easier/temping to do in telehealth settings) is a terrible system. "We may one day need to be able to verify if what AI wrote is real" is a terrible reason to change that.
Doctors (and increasingly patients) understand that a medical record can remain for your entire life. It will probably be seen by many different people within that time for valid reasons but medical records also get leaked/stolen/sold/illegally accessed. Patients need to be able to speak freely with their doctors and often depend on their discretion. Knowing that your every word will be recorded and kept in case somebody 10 years later has a question about what AI wrote in your file could keep people from being open and honest with their doctors.
> Except the industry (both the AI vendors and healthcare) are going YOLO¹ and relying on AI anyway.
Unless we get strong regulations to prevent it I'm afraid that you're right and that this is going to be a problem we experience in a lot of industries and areas besides healthcare. We see it happening in the justice system for example and it's already ruining people's lives.
We're in complete agreement here.
If we're not talking about an audio/video recording (a thing that nobody needs), the act of producing a record of a conversation involves choosing what goes into it.
We both agree that not every words that was said needs to go there. By far.
I guess it would be correct to say that there needs to be a record of every medical visit, but nobody needs a recording.
>Yes. This is what medical records are.
No. Medical records are limited extracts from conversations, which is your doctor and only your doctor is qualified to make, using "semantic analysis applied to your unique situation", not "linguistic probabilistic inference applied to conversation about your situation using token weights averaged over billion unrelated samples"
> It's not like the doctor is talking to you about which anime series are the best. You're talking about your health, your body, your disease, your treatment.
No jokes, no banter, no chit-chat, no complements to doctor's new Tesla?
> It's important to keep track of that.
Same fallacy Meta fell into when started tracking employees' keystrokes and mouse gestures. 90% of my mouse movements are just fidgeting, with no relation to the task at hand - and it is not a crime! But if I knew my mouse fidgeting is being watched, I'll make sure that percentage goes up to 99% - for the LLM which is gonna be trained off it to self-immolate over its NSFW nature.
I meant that these limited extracts do need to be recorded, that's all.
Read the rest of the comment :)
To get calibrated probabilities, you actually need to use calibration techniques, and it is extremely unclear if any frontier models are doing this (or even how calibration can be done effectively in fancy chain-of-thought + MoE models, and/or how to do this in RLVR and RLHF based training regimes). I suppose if you get into things like conformal prediction, you could ensure some calibration, but this is likely too computationally expensive and/or has other undesirable side-effects.
EDIT: Oh and also there are anomaly detection approaches, which attempt to identify when we are in outlier space based on various (e.g. distance) metrics based on the embeddings, but even getting actual probabilities here is tricky. This is why it is so hard to get models to say they "don't know" with any kind of statistical certainty, because that information isn't generally actually "there" in the model, in any clean sense.
It is too computationally expensive, which is why nobody does this for production inference. But there are alignment tools to extract out these latent-space probabilities for researchers in the frontier labs.
I'm pretty sure they are actively trained to avoid it.
Besides, like, what would you do if you asked your $200/mo AI something and it blanked on you?
> In this experiment, however, the model recognizes the injection before even mentioning the concept, indicating that its recognition took place internally.
But it then throws that distribution away / consumes it in the next token calculation. So it's not really tracking it per se.
Is it the token (or set of tokens) that are strictly > 50% probable or is it just the highest probability in a set of probabilities?
While generating bullshit is not ideal for a lot of use cases you don't want your premier chat bot to say "I don't know" to the general public half the time. The investment in these things requires wide adoption so they are always going to favour the "guesses".
I happen to believe that the flaw being discussed IS fundamental and inherent in the design and architecture of LLM - this is why I always put "AI" in scare quotes. I've spoken about it in some of my other comments, namely this https://news.ycombinator.com/item?id=47162553 and to some extent this https://news.ycombinator.com/item?id=48046333. And as you do, I, too, hope that I am wrong about the hype and its eventual clash with reality, but do not hold my breath.
Which both of us do.
If you ask me “how certain are you that the standard model of particle physics is true?” I’ll answer “I don’t know” because I don’t have any subject matter expertise, and philosophically I tend to hedge on questions like this anyway (“all models are wrong, some are useful”).
However, if you ask me “how certain are you that food is bland with no salt added, tastes better with some salt, but tastes bad with too much salt?” I would answer “very certain” because I have loads of direct experiments on this question in the kitchen. Furthermore, between these two extremes
To an LLM these are identical kinds of questions. All evidence has the same provenance: the training set. As of yet, we don’t have embodied AIs (robots) with multi-modal sensory inputs and online training. Until then, what we have remains a “brain in a vat fed on tokens” which, to me, is extremely weak from an epistemic perspective.
As far as I understand it, the various probability matrices boil down to: what token has the highest likelihood of coming next, given this set of input tokens. Which then all gets chucked away and rebuilt when the most likely token is appended to the input set.
Objective assessment of internal state - again, to my non-expert eye - doesn’t appear to have any way to surface to me.
Big-if my rough working understand is more or less correct - your calibration point makes a lot of sense to me. I’m not sure that it would make sense to someone who eg considers some form of active thinking process that is intellectualising about whether to output this or that token.
I'm not sure who is doing what training exactly, but I can say that (inconsistently!) some of my attempts to get it to solve problems that have not yet actually been solved, e.g. the Collatz conjecture, have it saying it doesn't know how to solve the problem.
Other times it absolutely makes stuff up; fortunately for me, my personality includes actually testing what it says, so I didn't fall into the sycophantic honey trap and take it seriously when it agreed with my shower thoughts, and definitely didn't listen when it identified a close-up photo of some solanum nigrum growing next to my tomatoes as being also tomatoes.
> Besides, like, what would you do if you asked your $200/mo AI something and it blanked on you?
I'd rather it said "IDK" than made some stuff up. Them making stuff up is, as we have seen from various news stories about AI, dangerous.
They really aren't, especially if you consider the chain of thought / recursive application case, and also that you can't even assume e.g. a difference of 0.1 in softmax values means the same relative difference from input to input, or that e.g. an 0.9 is always "extremely confident", and etc. You really have no idea unless you are testing the calibration explicitly on calibration data.
> But there are alignment tools to extract out these latent-space probabilities for researchers in the frontier labs
You can get embeddings: if you can get calibrated probabilities, you'll need to provide a citation, because this would be a huge deal for all sorts of applications.
And now I'm certain we're taking past each other. I'm not talking about calibrated probabilities at all. Just the notion of "how confident do I feel about this?" which is what I interpreted the question above to be about. You can get that out of an LLM, with some work.
There is nothing straightforward about this, and no, there is no such formula.
> I'm not talking about calibrated probabilities at all. Just the notion of "how confident do I feel about this?"
If all you care about is vibes / feels, sure. If you actually need numerical guarantees and quantitative estimates to make your "feelings" about confidence mean something to rigorously justify decisions, you need calibration. If you aren't talking about calibration in these discussions, you are missing probably the most core technical concept that addresses these issues seriously.
Finding a way to objectively calibrate a sense of "how confident do I feel about this?" would be fantastic. But let's not move goal posts. It would still be incredibly useful to have a machine that can merely matches the equivalent statement of confidence or uncertainty that a human would assign to their mental model, even if badly calibrated.
> It would still be incredibly useful to have a machine that can merely matches the equivalent statement of confidence or uncertainty that a human would assign to their mental model, even if badly calibrated.
If human feelings are badly calibrated, they are useless here too, so no, I don't agree. Things like "confidence" only matter if they are actually tied to real outcomes in a consistent way, and that means calibration.