Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Unexpected responses from ChatGPT: Incident Report

status.openai.com

Post reply on HN
Print view View original

191–200 of 277 posts

Re: Unexpected responses from ChatGPT: Incident Report

#191

Post by renonce »

iforgotpassword wrote: post #161
montroser wrote: post #26

This explanation feels unsatisfying. It's so high-level that it's mostly void of any actual information. What was the wrong assumption that the code made that caused this wrong behavior? Why was it not caught in the many layers of automated testing before it made its way to production? What process and procedural changes are being implemented to reduce the risk of this class of bug happening again? Presumably all of…

To be honest > On February 20, 2024, an optimization to the user experience At that point, about 10 words in, I already wanted to stop reading because it starts with the "we only wanted the best for our customers" bullshit newspeak. Anyone else going off on that stuff too? I'm pretty much already conditioned to expect whatever company is messaging me that way to take away some feature, increase pricing, or otherwise…

That sounds like a good use case for GPT. A GPT that automatically highlights such corporate speak and hints “WARNING: bullshit ahead”. I’m 100% sure it’s technically very easy to engineer such a model.

Do you think OpenAI’s superalignment will ever allow you to make such a model?

Re: Unexpected responses from ChatGPT: Incident Report

#192

Post by troupo »

okdood64 wrote: post #10

Say what you will about Google, but bugs such as these, released to the public (or their enterprise customers) this debilitating to a core product, are exceedingly rare.

I see you haven't seen the Gemini debacle: https://twitter.com/alexthechick/status/1760331428133253312

Re: Unexpected responses from ChatGPT: Incident Report

#193

Post by skissane »

bongodongobob wrote: post #156
skissane wrote: post #139

Earlier quoted context omitted.

> Saying rocks are conscious is a poor summary > The whole thing is a continuum and everything is made of matter, so there's a little bit of potential consciousness in all matter I disagree that my summary is "poor"–because some panpsychists do say that rocks are actually conscious individuals, as opposed to merely containing "a little bit of potential consciousness". The IEP article I linked contains this quote from…

I think it would be more fitting to say a rock, a grain of sand, or an atom is the least conscious thing. Pansychism isn't trying to claim it can think or anything like that. If that's your takeaway I think you are misunderstanding the analogy. A rock exists in the world and exerts it's "will" granted to it by the laws of physics. So rolling falling etc is how it "interacts" with its world. You can think of our brain…

> I think it would be more fitting to say a rock, a grain of sand, or an atom is the least conscious thing. Pansychism isn't trying to claim it can think or anything like that. If that's your takeaway I think you are misunderstanding the analogy.

No, I don't think I've misunderstood anything; on the contrary, I think you are mixing up thought and consciousness. Thought is just one type of consciousness among many; to say something is conscious doesn't necessarily imply it can think–possibly, many things are conscious yet unable to think thoughts (whether verbal thoughts, or thoughts composed of mental images, or whatever). I'd say that to be conscious is to be an entity for which Thomas Nagel's famous question "What is it like to be a bat?" has an answer; it is possible that the question has an answer for an entity, yet it is incapable of thought. A being could be completely conscious yet also completely incapable of thinking anything.

Re: Unexpected responses from ChatGPT: Incident Report

#195

Post by cfn »

montroser wrote: post #26

This explanation feels unsatisfying. It's so high-level that it's mostly void of any actual information. What was the wrong assumption that the code made that caused this wrong behavior? Why was it not caught in the many layers of automated testing before it made its way to production? What process and procedural changes are being implemented to reduce the risk of this class of bug happening again? Presumably all of…

I don't see the issue, in this type of software a one off bug in tokenization, for example, would create the type of nonsense people saw. That would make sense given their high level explanation.

Re: Unexpected responses from ChatGPT: Incident Report

#196

Post by skissane »

card_zero wrote: post #185
skissane wrote: post #140

Earlier quoted context omitted.

> I guess an obvious objection is "why?" Because coming up with criteria for determining what is and isn't conscious, and justifying the particular criterion you choose (as opposed to all the alternatives) is hard . Faced with the difficulty of that problem, the two simplest solutions are the extremes of panpsychism (everything is conscious) and eliminativism (nothing is conscious)–since for both the "criterion of co…

Parsimony leads to wrangling about which is the simplest explanation, yes. Consciousness is ill-defined, also. It does however seem to stop when the brain is destroyed, which makes it unlikely to be present in fundamental particles, or in parts of a disintegrated brain.

> It does however seem to stop when the brain is destroyed, which makes it unlikely to be present in fundamental particles, or in parts of a disintegrated brain

Does consciousness really "seem to stop" when the brain is destroyed? We can't directly observe anyone's consciousness other than our own. It is true that, when people die, we cease to have access to the outward signs we use to infer they are conscious – but we don't know whether that's because their consciousness has ceased, or whether that is because it has "gone somewhere else" to which we don't currently have access. We never have direct access to the inner reality of consciousness for anyone other than ourselves, and we can only know the temporary cessation of our own consciousness, its permanent cessation (if that is possible) is in principle unknowable to us.

Speaking of our consciousness "going somewhere else" at death (an afterlife) doesn't have to be a religious claim – there are heaps of "naturalistic" ways in which consciousness might survive death (e.g. Boltzmann brains, computer simulations, quantum immortality). Since we don't know whether any of those ways actually happen, we don't know whether or not consciousness actually stops at death.

Re: Unexpected responses from ChatGPT: Incident Report

#197

Post by TaylorAlexander »

Grimblewald wrote: post #100
atleastoptimal wrote: post #17

Almost a tautological PR paragraph: "The words were wrong because the model chose the wrong words"

The missile knows where it is https://www.youtube.com/watch?v=bZe5J8SVCYQ

The missile is eepy

https://youtu.be/Csp_OABIsBM

Re: Unexpected responses from ChatGPT: Incident Report

#198

Post by prmoustache »

tomas789 wrote: post #174
bjackman wrote: post #171

Earlier quoted context omitted.

I don't a the problem with this communication, except they just shouldn't call it a "postmortem". In a medical setting this is equivalent to saying "the patient was coughing up blood, and now they're dead". It's relaying some potentially useful info, and I guess it's _technically_ a post-mortem analysis but it doesn't fit expectations for the phrase!

It is an industry standard word by now.

The standard imply analysis and measures taken for similar issues to not happen.

Re: Unexpected responses from ChatGPT: Incident Report

#199

Post by kruxigt »

p10_user wrote: post #99
toddmorey wrote: post #48

Earlier quoted context omitted.

I mean, this is almost poetic: Classic plays, wholly told. Shadow into form, hands as cue. Keep it at the dial, in right on Pitch.

ChatGPT seems to appreciate it https://chat.openai.com/share/37a490fe-e7b4-4042-a560-59b519...

[dead]

Re: Unexpected responses from ChatGPT: Incident Report

#200

Post by prmoustache »

BeefySwain wrote: post #24

I am used to postmortems posted to here being a rare chance for us to take a peek behind the curtain and get a glimpse into things like architecture, monitoring systems, disaster recovery processes, "blameless culture", etc for large software service companies. In contrast, I feel like like the greatest insight that could be gleaned from this post is that OpenAI uses GPU's.

They could have said "shit happened" and it would have been as informative tbh.
Post reply on HN