Original sources aren’t necessarily infallible either, but that’s beside the point, asking ChatGPT needs to be viewed as asking a person to recount something from memory. It hasn’t been reviewed or edited so it’s like a first draft rather than a published document. That’s the error that people tend to make since they treat it like a search engine instead of a conversation.
They just want to reinforce their own bias that OpenAI BAD and DUMB, rationalism GOOD! When it’s their own fault for not understanding enough theory to know what happens if there were to be a loss of precision in the predicted embedding vector that maps to a token. If enough decimal places are lopped off or nudged then that moves the predicted vector slightly away from where it should be and you get a nearby token instead. Instant aphasia. The report said it was a GPU configuration problem so my guess is the wrong precision was used in some number of GPUs but I have no idea how they configure their cluster so take that with a giant grain of salt.
Agreed, but I think they raise a fair point that some kind of automated testing probably should've existed to look for sudden major changes in prompt -> output production between versions, with lots of prompt test cases. Maybe this testing did exist but this problem just didn't surface, for some reason?
Yuck. So many half truths and outright lies. This is straight up propaganda. Look into any of the assertions given here and you’ll see what I mean. Designed to fool headline readers only.
Honestly, I don’t think the unemployment rate will change much. Humans are great at inventing things to do and if other people see those things as valuable they will pay for them. I do think the world will look very different, maybe even unrecognizable but it’s not going to be full of people doing nothing.
In what way does anyone have a monopoly on generated images and video? Last I checked there were several major players and more startups than you can shake a stick at.
It won’t last. There’s a massive incentive to build more GPUs and develop specialized chips and everyone who can is scrambling to meet that demand. The technology is not some trade secret that no one can copy which is why there are so many people and companies diving into this market now. Hardware is a bit slow to ramp up production of but it will get there eventually because there’s money to be made.
Does that matter when the models they generate are given away for free?
You can make your argument validly against DALL•E or Midjourney families, but we've also got the Stable Diffusion family of models that anyone can just grab a copy of.
I’m talking about generative ai VS human artists. But in this case it seems like OpenAI specifically has a massive leap over everyone else with this video generation. So whether they have a monopoly over that remains to be seen.
What does not remain to be seen though is that generative ai is going to put a lot of artists out of work.
You can argue about the good and bad of that but it’s defo happening.
I’m pretty sure you could record and transcribe in person meetings just as easily. I’d bet there are several services doing that now but it would also be an incredibly simple project to build in house. The only problem is identifying who is speaking which could be a deal breaker for some uses.
>I’m pretty sure you could record and transcribe in person meetings just as easily.
The issue is that in person you constantly have informal "meetings" where information is shared and impossible to record, unless you force everyone to walk around with sound recorders on. Which come to think of it is something that might well be done in some companies I've worked for before.
>The only problem is identifying who is speaking which could be a deal breaker for some uses.
Trivial, it's a problem that's been solved for a while and I managed to get it working in a couple of days of messing about.
>The issue is that in person you constantly have informal "meetings" where information is shared and impossible to record
I'm positive that people working remotely can just as easily have informal meetings were they won't be recorded. In fact, I'm absolutely sure that a good segment of the workforce will make sure they have meetings in that informal sense, specifically so the recordings aren't happening.
While I do think we’ll go through a phase where AI writes a significant amount of code I think it will eventually replace the need for most software. To illustrate what I mean, take your browser example. Why would you even want a browser when you can just make a request to the AI and it can give you exactly the information or images you are looking for formatted in your preferred style and edited and annotated to perfectly match your ability to comprehend it? Honestly, I think it will go beyond even that. Our current mode of consuming information may be entirely replaced by simply conversing with personalized AI. We won’t search for information, our assistant will spontaneously tell us what we need to know and we will interrogate it about things that spark our interest. There will be no UI, only an active agent that you speak with and write to who answers your questions, acts on your behalf and delivers the information you want as a document on a tablet, a video on a screen, an immersive VR experience or performed as ballad if the mood strikes you.
There’s no chance they give you full reigns of output considering present business paradigms at least. There’d be no differentiating your product from competitors or justifying upgrades if it was always how you wanted it to work.
What you are describing is an traceable transformation of code in which there are several intermediate layers that people can inspect and understand. They can inspect the exact rules for that transformation. The process is repeatable and verifiable.
What I am describing is a black-box stochastic generation of low level code in which there is no higher level representation anymore. AI generating Assembly not by a set of rules, but using statistics. There will be no individual layers to unwind or inspect, because for AI it doesn't need them. Our separation of concerns was built for our human brains and limiting complexity of projects to our understanding.
If you have an AI capable of writing machine code based on natural language you likely also have an AI that can translate that machine code to any other language you would like. You could then use a normal compiler to verify if it is correct and then read the code yourself. Or you could just get good at reading and writing assembly.
Yes, that is exactly my point. You will have to rely on AI to translate it back out, but that translation is built on probabilities not machine rule-based translation. So you can ask and have the AI explain everything to you, but you are still trusting the "black box" to tell you what is happening. Very different from today.
Also, you can "get good" at reading assembly, but that doesn't matter if the AI can output a custom OS from scratch and a custom VM to execute the program it wrote to solve your use case. It will be so impossibly complex that it would be the equivalent studying protein folding.
Instead people will just trust the AI.
It also won't help you if the code base the AI produces for a SaaS app is a million lines of assembly.
Instead of having different layers of OS, compiler, high level language, an AI will just be able to produce one layer. because after decades of trusting the AI to write our code, why wouldn't it?
The current gen of AI outputting code that in human-centric programming languages will be a blip in the history of AI. As it advances, it can just skip that step.
Its will be orders of magnitude more complex and opaque than anything we have today.
Pomodoro never worked for me since I work best when I get into flow and keep going for 1 to 4 hours. What I have found helps is planning as much as possible up front and dedicating time to do so. I braindump everything then when that’s exhausted I rewrite it in a structured way and add detail, breaking it up into sections and gathering resources I need, then I make an exhaustive todo list of tiny steps in chronological order and I mean really tiny steps with concrete outcomes. Then when I start working I just follow the plan and since I’m very familiar with it by this point I don’t have to refer back to it all the time but I do make sure to check it in order to stay on track usually when I reach a point where I’m uncertain of what to do next. I also add to the document when necessary.
