> And if these rumors, that GPT-4 is inherently un-deterministic and unreliable, are true then most enterprises are better off finetuning open source LLMs—which are just as capable
Wait, am I misunderstanding you? I feel like I've had a head injury or something, because I've never heard of an open source LLM that's as capable as GPT-4 (in most scenarios).
I guess they could bet heavily against themselves, try to get prediction markets to coalesce around the result being genuine, them come out saying "oops, we were mistaken"? Bam, rich!
For what it’s worth, I’m 41 and I often prefer video content. It keeps me on track when my attention span would get in the way were I reading a longer tutorial or manual.
I don’t know how to square this with the fact that I routinely ask GPT-4 for code and get back answers that are in fact correct. You’re making it sound like it’s purely random, like I’m just getting back a string of zhrueksnbd8373&;! or something.
My girlfriend has a rather uncommon name, but her FirstNameMiddleName@gmail.com gets emails meant for someone else rather frequently. Sometimes it’s even official PII or PHI. She’s tried to let people know but they’re surprisingly sometimes not all that receptive? So I think she ignores it now.
I have @CommonFirstNameCommonLastName on Twitter, and I occasionally get mentioned or DMed by people thinking of an entirely different person. It’s not often enough to be too annoying, though.
There’s a history of actual usage among some subsets of people, not for everyone. Regardless, it doesn’t really matter if there’s history as long as usage is in fact changing. “Literally” meant one thing much longer than it has been a superlative.
I don’t know that Google has said exactly how it’s implemented this feature under the hood. The underlying Palm 2 model does not natively handle images, though, so it’s relying on other services like Lens to handle the image understanding part. I suspect these are feeding imperfect information back to Bard, who relays that info to the user.
You say anyone interpreting your earlier statement as being about those who agree with Google lack basic logic, yet you repeatedly say things like this to people defending Google:
> And given that I'm not a pathetic bootlicker, I don't immediately look to rationalize for them.
If you don’t want to imply that you think this about anyone taking Google’s side, maybe use some more nuanced language. Just a suggestion.
Frankly, it seems to me that you want it both ways — to insult people who disagree with you but then hide behind some purely technical sense in which you didn’t.
This is probably because it can’t actually understand images, it’s relying on other services to deal with it. So it can do an image search to find what something might be an image of, or use ocr to extract text, but it’ll fail on tasks that involve the idiosyncrasies of each image.
> I mean things like it refuses to speculate about what medical condition might cause some given symptoms.
Are you sure about this specifically? I’ve recently had zero trouble getting GPT-4 to give potential diagnoses for a given set of symptoms, though perhaps it’s an issue of prompting.
Great. And if I want your opinion on how smart it is to be a bit silly with LLMs when I’m not even attempting to test them systematically, I’ll be sure to ask you.
Not exactly. The quote about highway collisions referred to a time in the future where there are substantially more satellites in orbit. If it’s true that these maneuvers are increasing exponentially, that could be a genuine problem.
This is hysteria. Mozilla's "political action" is minimal compared to their browser work. Pointing fingers at mostly nonexistent institutional wokeness as the reason for their low market share ignores the many true reasons (including network effects and consumer inertia with a "good enough" product that everyone already knows how to use).
> that's... not a valid argument for this discussion?
I believe the intention was to argue against the notion that 95% of users don't read release notes. The contention seems to be that, because Firefox users are more technical (or enthusiast, perhaps), more than 5% of Firefox users would read be aware of this change.
I still use Facebook for the family connections and the pet cat/dog/rabbit accounts I follow, but I'm essentially done with Twitter now. I was already not posting there basically at all, and the rate limiting means I won't be reading it, either. Threads couldn't be released at a better time for Meta.
This is a great point. There’s a lot of wooly thinking about what it means for a system to be intelligent, particularly “super” intelligent. People seem to think we’ll create a machine that is almost literally infallible — able to predict both physical systems and human behavior with perfect foresight many steps in advance. I’m not sure that’s even possible, let alone likely.
The analogy is just that: an analogy, and a very imperfect, misleading one. The working of the brain may have motivated early research, but GPT (as instantiated in hardware) does not operate or learn in a way similar to a human brain.
Intent only goes so far. If I continually but unintentionally reproduce copyrighted works verbatim, I could still face consequences, particularly if I did not show due diligence in preventing it from happening in the first place.
Sorry, I was being a little flip. There's more to it than that, of course. Is the performance sufficiently transformative, is it educational, is it non-profit, etc.
> Should I have gotten help from the government to instead stick to old style sysadmin work?
I don't think anyone beyond a few marginal voices are calling for a ban on job automation. What they seem to prefer is that, if they are to be automated out of a job, they should be compensated for their copyrighted works having been used in the process of doing so.
Regardless, at the very least people who are being automated should get some government support. Not everyone can easily retrain.