Assuming that people only share conversations they think are good, would that be bad? Isn’t that the basis of RHLF?
There are a few times on Reddit that I want to explain something that I know well. But it will be a long post.
I’ll be lazy and ask ChatGPT the question, either verify it’s correct based on what I know, ask it to verify its answer on the web - the paid version has had web search for over year - or guide it to the correct answer if I notice something is incorrect.
Then I’ll share the conversation as the answer and tell the poster to read through the entire conversation and tell them that I didn’t just naively ask ChatGPT. It will be obvious from my chat session.
I’ve had pretty good luck when having it write Python automation scripts around AWS using Boto3.
If it’s a newer API that ChatGPT isn’t trained on, I would either tell it where to find the newest documentation for the API on the web or paste the documentation in.
It usually worked pretty well.
If the author of the library wrote good documentation and sample code, you wouldn’t need StackOverflow hypothetically if ChatGPT was trained on it
Apple is training its own autocomplete for Swift on its documentation and its own sample code.
We don't have to guess. Just look at languages which have been around for a while, achieved some baseline level of popularity to have a decent amount of public code available, like Elixir.
I haven't found an LLM that could reliably produce syntactically correct code, much less logically correct code.
Since LLMs have been a thing, I’ve been heavily involved in the AWS ecosystem and automation.
ChatGPT is well trained on the AWS SDK for various languages. I can usually ask it to do something that would be around up to a 100-200 line Python script and it gets it correct. Especially once it got web search capabilities, I could tell it to “verify Boto3 (the AWS SDK for Python) functions on the web”.
I’ve also used it to convert some of my handwritten AWs SDK based scripts between languages depending on the preferences of the client - C#, Python, JavaScript and Java.
It also does pretty well at converting CloudFormation to idiomatic CDK and Terraform.
I was going into one project to teach the client how to create deployment pipelines for Java based apps to Lambda, EC2 and ECS (AWS’s Docker orchestration service).
I didn’t want to use their code for the proof of concept/MVP. But I did want to deploy a sample Java API. I hadn’t touch Java in over 20 years. I was a C#/Python/Node/(barely) Go developer.
I used ChatGPT to create a sample CRUD API in Java that connected to a database. It worked perfectly. I also asked about proper directory structure.
It didn’t work perfectly with helping me build the Docker container. But it did help.
On another note: it’s not too much of leap to see how Visual Studio or ReSharper could integrate an LLM better and with static language, guarantee that the
code is at least syntactically correct and the functions that are call exist in the standard library or the solution.
They can already do quick, real time warnings and errors if your code won’t compile as you type.
Its own answers, with feedback about whether the answers seem to have worked.
Learning to predict what word will lead to a successful solution (rather than just looking like existing speech) may prove to be a richer dataset than SO originally was.
Most of the time this would happen in the form of an interactive debugging session, with immediate feedback.
Code review is its own domain. In general at some point LLMs need to be trained with a self-evaluation loop. Currently their training data contains a lot of "smart and knowledgeable human tries to explain things". And they average out to conversation that is "smart and knowledgeable...about everything". That won't get us to, "Recognizably thinks of things that no human would have." For that we need to get it producing content that is recognizably higher than human quality.
For that we should find ways to optimize existing models for an evaluation function that says, "Will do really well on self-review." Then it can learn to not just give answers that help with interactive debugging, but actually give answers that will also do well with more strenuous code review. Which it taught itself how to do in a similar way to how AlphaZero manages to teach itself game strategies.
Most of that doesn't deserve to be called an improvement. That message encryption is one of the worst designs I've ever seen. The algorithm isn't open enough that anybody can see if what it's doing matches the code.
And instead of censoring "secretly" in partnership with the US government they do so in partnership with the Indian government as well as individual Republican politicians.
> Most of that doesn't deserve to be called an improvement.
Hard disagree. The platform is objectively much better than it was under Jack Dorsey.
> The algorithm isn't open enough that anybody can see if what it's doing matches the code.
Fair, but more is available than before.
> ...in partnership with the Indian government as well as individual Republican politicians.
Would you prefer X break India's laws, specifically Section 69A of the IT Act? I suppose they could do that and then the Indian government would shut them down in India.
Also, are you talking about X suspending Ken Klippenstein's account for doxxing JD Vance? X's rules explicitly state that doxxing people will result in a suspension, so maybe Klippenstein should have thought of that before posting Vance's home address and most of his social security number.
It's used in a not particularly advanced way - bluesky uses repositories per each user for all their public social activity (posts, likes, etc) and it creates a Merkle tree as an index for them and signs that index (this enables stuff like authenticated content addressing and efficient verification).
SQLite simply stores those posts and that signed Merkle tree. The PDS account host server also has another SQLite DB with a list of its accounts.
They have fancier stuff in their appview server and relay server
It's currently built into the applications, not configurable by default. It's preferable if everybody in a thread / space / community uses the same one because otherwise differences can cause validation failures and conflict and thus break threads, resulting in that people can't follow many discussions, thus applications don't really expose it.
It's the kind of thing where it's preferable that either everybody switch at once, or that new applications with different "lexicon" (post types / social media format) picks a new default from the start.
That's too bad, it made me think if there were easy to switch user perferences then it would be truly decentralized..
Is there a way to clone the app and make this easy to find as a button to switch?
(unfamiliar with the licensing and the app in general, some things I have seen make me excited that it's possible to be uncensored and decentralized, then things like this come up and I wonder if I should put time into it or something else to promote)
Everything is open source so yes you can, it's mostly a question of practicality.
There's already third party clients, account hosting servers, etc, as well as different apps building on the same system (and which can use the same accounts and data store!) like blogs and more. Most devs are trying to coordinate their custom extensions so it doesn't cause conflict.
If it weren't coordinated we'd easily end up in the same place as Mastodon with their spurious server blocks where large parts of conversations are broken for most users.
I'd love more info about "Mastodon with their spurious server blocks where large parts of conversations are broken for most users."
- who and why that happens and what it looks like to the different users that have servers blocked or what not.. is it clear they are missing messages that are from blocked servers?
The irony is that what they claim happened is instead what would have happened only if the case was resolved in the opposite direction. Everybody having to pay for API licenses for absolutely everything would be disastrous. The gridlock would be insane.
As others has said elsewhere, the jury rules on facts and judges rules on law. SCOTUS are judges. Under one interpretation of law, the most recent jury findings of fact held that Google infringed.
SCOTUS altered the interpretation of the law, thus removing the legal justification for why the jury found infringement. Since they did not alter any findings of fact (they did not need to) this ruling is legally fine.
A jury can say you definitely did X, that court's judge can say X is illegal, and SCOTUS can then declare X is legal so it doesn't matter anymore if you did X.
