This is the advent of something really earth-shattering.
And I say that as a skeptic of most of these deep learning things.
- opengl raytracer with compilation instructions for macos
- tictactoe in 3D
- bitorrent peer handshake in Go from a paragraph in the RFC
- http server in go with /user, /session, and /status endpoints from an english description
- protocol buffer product configuration from a paragraph english description
- pytorch script for classifying credit card transactions into expense accounts and instructions to import the output into quickbooks
- quota management API implemented as a bidirectional streaming grpc service
- pytorch neural network with a particular shape, number of input classes, output classes, activation function, etc.
- IO scheduler using token bucket rate limiting
- analyze the strengths/weaknesses of algorithms for 2 player zero sum games
- compare david hume and immanuel kant's thoughts on knowledge
- describe how critics received george orwell's work during his lifetime
- christmas present recommendations for a relative given a description of their interests
- poems about anything. love. cats. you name it.
Knowing that code is correct is as important as the code itself, and this is why we do code review, write tests, have QA processes, use logging and observability tools, etc. Of course the place that catches the most bugs is the human writing the code, as they write it.
This feels like a nice extension to Copilot/etc, but I’m not sure it’s as general as people think.
Perhaps an interesting challenge to pose to it is: here’s 10k lines and a stack trace, what’s the bug. Or here’s a database schema, what issues might occur in production using this?
Once I trust the tests, I generally trust the code.
I've seen Copilot generate code I read and thought was correct, that went through code review and everyone thought was correct, that had tests written for it (that nearly covered everything), and that even when it failed, was hard to spot the issue.
It turned out it got a condition the wrong way around, but given the nesting of conditionals it wasn't obvious.
I don't think a human who was thinking through the problem would have made the same mistake at the point of writing, in fact I think that the mind state while actually writing the code is hard to reproduce at any later time, which is why code review isn't great at catching bugs like this.
Ah must be a Spring application ...
This seems like the lowest number that would be useful. Below that it's not really a problem to debug, but at that point there's typically enough complexity that some help would be useful as you forget edge cases and features in the codebase.
For demonstration purposes doing it with 100 lines might be ok, but for professional use it kinda needs to understand quite a lot! Like a minimum of that order of magnitude, but potentially millions of lines.
FWIW, I've never used Spring. My experience is mostly Django, iOS, non-Spring Java, and some Android.
That's against the EULA if OpenAI may want to make a similar model:
> (iii) use the Services to develop foundation models or other large scale models that compete with OpenAI;
https://openai.com/api/policies/terms/
Seems to be about developing models and not just restricting you from training them with it.
Kind of ironic given that OpenAI builds and trains all of their models on stuff they "found" in the open.
Either everything is fair game for training, or nothing at all is.
If I were a judge ruling on this matter, I would absolutely rule that bootstrapping a model from OpenAI outputs is no different than OpenAI collecting training data from artists and writers around the web. Learning is learning.
Might be worth trying to use the outputs to bootstrap. What are they going to do about it? Better to ask forgiveness until the law is settled.
They may not need to.
I love these sorts of loopholes. OpenAI is actively trying to curb the potential of their AI. They know how powerful it is. Being able to see a taste of that power is endlessly exciting.
Code-completion > Abstraction.
Also there are specific techniques for validating that you are model training procedure is directionally correct, such as generating a simulated data set and training your model on that.
The difference is that now, you are more of a code reviewer and editor. You don't have to sit there and figure out the library interface and type out every single line.
- Implement a simple ray tracer in C++ using opengl. Provide compilation instructions for macos.
- Create a two layer fully connected neural network with a softmax activation function. Use pytorch.
- Implement the wire protocol described below in Go. The peer wire protocol consists of a handshake followed by a never-ending stream of length-prefixed messages. The handshake starts with character ninteen (decimal) followed by the string 'BitTorrent protocol'. The leading character is a length prefix, put there in the hope that other new protocols may do the same and thus be trivially distinguishable from each other.
- We are trying to classify the expense account of credit card transactions. Each transaction has an ID, a date, a merchant, a description, and an amount. Use a pytorch logistic regression to classify the transactions based on test data. Save the result to a CSV file.
- We are configuring settings for a product. We support three products: slow, medium, and fast. For each product, we support a large number of machines. For each machine, we need to configure performance limits and a mode. The performance limits include iops and throughput. The mode mode can be simplex or duplex. Write a protocol buffer for the configuration. Use an enum for the mode.
- How were George Orwell's works received during his lifetime?I asked for jokes this morning and initially it made excuses and wouldn't give me jokes until I tweaked the prompt.
Later I refreshed the chat and pasted in the original prompt and got jokes right away, with no excuses.
(I was asking for jokes on the topic of the Elon Musk Twitter acquisition. My personal favorite: "With Elon Musk in charge, Twitter is sure to become the most innovative and futuristic social media platform around.")
[1] https://ai.googleblog.com/2022/01/lamda-towards-safe-grounde...
Perhaps it is even prohibitively expensive.
The number of people using Google Search is easily 1000x larger than the number of people using OpenAI, if not more.
I’ll leave it for you to decide which is the pessimistic option.
(Google project pitchfork.)
"GPT-5, write me a poem."
"Roses are red, violets are blue, sugar is sweet, and this output is sponsored by Raid Shadow Legends"
Training the models weren't that expensive either, and that's a one time deal.
If we couldn't tell that it was cheaper to use renewable over fossil fuels or eat chicken over beef as two examples there could be big efficiency and allocation problems.
AIs with total surveillance perhaps? Would need weather-control too or giant surpluses squirreled away in a distributed fashion.
Some kind of ascetic commune, but at global scale? Might work in a military setup when fighting the alien invasion.
“Write a YouTube script sponsored by NordVPN by Littlefinger about the invasion of Westeros by Daenerys.”
“ Hi there! I'm Littlefinger and I'm here to talk to you about the recent invasion of Westeros by Daenerys. As you may know, Daenerys is a powerful queen from the east who has been making her way across the land with her army and her dragons.
Many people are worried about what this means for the future of Westeros, and for good reason. Daenerys has a reputation for being a fierce and unstoppable force, and she has already proven herself to be a formidable enemy.
But there is one thing that can help protect you from the dangers of Daenerys and her invasion: NordVPN. NordVPN is a virtual private network service that allows you to securely browse the internet and keep your online activity private.
With NordVPN, you can protect yourself from being targeted by Daenerys and her forces. Whether you're trying to keep your personal information safe or you just want to make sure that your online activity remains private, NordVPN has you covered.
So why not give NordVPN a try today? With its easy-to-use interface and powerful security features, NordVPN is the perfect tool to help you stay safe during these uncertain times. Plus, with its 30-day money-back guarantee, you have nothing to lose!
Thank you for watching, and stay safe with NordVPN.”
https://twitter.com/goodside/status/1599094067534516225?t=n9...
And there will be companies offering the product for free in order to grab that market share and steal users from the paying ones.
That's how the internet companies strived since ~1998
Spoken as if a sustainable and profitable business model prevents ads from being shoved into a product. It does not.
here is a function in python3
def foobar(i):
j = i + 1
if j > 10:
print("yam")
else:
print("baloney")
change the function foobar so that foobar(9) also prints "yam"
---
To change the behavior of the function so that foobar(9) also prints "yam", you can modify the condition in the if statement to check if j >= 10 instead of j > 10. This will cause the if block to be executed when j is equal to 10, as well as when it is greater than 10.
Here is the modified function:
def foobar(i):
j = i + 1
if j >= 10:
print("yam")
else:
print("baloney")
foobar(9) # prints "yam"
foobar(10) # prints "yam"
At some point when either using or running projects around prompt engineering we have to switch from our high modernist viewpoint which is baked into our culture. And instead, deal with it explicitly as a statistical model. And even when such a model is super smart, know by the way it is constructed that it is fuzzy and often "wrong" (like humans are).
baloney
yam> change the function foobar so that foobar(9) also prints "yam"
It interpreted the prompt in the same way that I would, making foobar(9) only print "yam", but I guess I can see the interpretation where someone might think the prompt was asking to print both. It brings awareness to the effect of cultural bias in the training set and how that might require some people with different ways of speaking to add more clarification to get the output they want.
By your logic, a prompt like "Print foo" could result in "foo bar". You didn't ask for "bar" but the output is still valid according to you.
2. Change ... so that foobar(9) also prints "yam"
1 is not ambiguous, but 2 is. It could mean that it does it's original behavior and also does the new behavior. But it could mean like foobar(10), foobar(9) also prints only "yam" as you are interpreting it to be.
I could imagine that the whole point was to see what GPT-3 did with the ambiguity, and it didn't even acknowledge it. Given that the prompt is already ambiguous I'm not sure why you'd assume that there would have to be a single expectation of the order of the prints, either one should suffice.
Or maybe it was just a mistake in the prompt because English is hard. Certainly based on the code, the assumption that GPT-3 made is the most natural one.
But seems within the realm of possibility -- with better training and better prompting.
The potential to summarize a topic and then generate a PPT output is tantalizingly close :)