231–240 of 412 posts
Microsoft would need to pay me to use Copilot. Seems like a major scam for them to learn from our code and then tell us we can’t use it to make our own competing AI systems. “Limits on use of data from the AI Services. You may not use the AI services, or data from the AI services, to create, train, or improve (directly or indirectly) any other AI service.” From https://www.microsoft.com/en/servicesagreement#13r_AISer…
Has anyone compared the updated Copilot with Cursor? The main updates I am wondering about are model selection and multi-file edits. I used copilot before these features, changed to Cursor and now I am wondering how much Copilot has closed the gap.
The UI is still changing slightly every couple weeks as they improve things and polish it, but it's become a big enough part of my day to day that it's pretty much always open on the right pane of vscode for me.
I have been using Cursor, but I am a vim user at heart. What is the best plugin I can use that gives us the power of AI coding, but for vim? Every time I look, the plugins seem to be far behind Cursor.
The real cost is that society at large is no longer contributing to the StackOverflow, so problems and solutions are all now stored in proprietary databases (which granted SO also was) but also now stored in an invisible proprietary database
StackOverflow's database was public and shared at the Internet Archive until recently: https://archive.org/details/stackexchange
They've now moved it onto their own infra, ostensibly so people have to agree not to use it for LLM training: https://meta.stackexchange.com/questions/401324/announcing-a...
Microsoft would need to pay me to use Copilot. Seems like a major scam for them to learn from our code and then tell us we can’t use it to make our own competing AI systems. “Limits on use of data from the AI Services. You may not use the AI services, or data from the AI services, to create, train, or improve (directly or indirectly) any other AI service.” From https://www.microsoft.com/en/servicesagreement#13r_AISer…
They claim it's fair use for them to steal all data they want, but you're not allowed to use AI data output, despite this data literally not being subject to copyright protections on account of lacking a human author.
And especially Github. They already have an enormous corpus that is licensed under MIT/equivalent licenses, explicitly permitting them to do this AI nonsense. All they had to do was use only the code they were allowed to use, maybe put up an attribution page listing all the repos used, and nobody would've minded because of the explicit opt-in given ahead of time.
But no. They couldn't bother with even that little respect.
I ditched Copilot for Cursor and never looked back. Cursor might be the only other AI product with product market fit than ChatGPT, it's that good
Why do I say this? Because their cost to deliver their product exceeds their revenue. That tells me the product (as it is today) does not match the market demand.
Earlier quoted context omitted.
1. I believe AI is detrimental. It makes us go too fast. It's all about production now, pure efficiency over individuals. 2. AI is too dangerous. Whatever innovations used for benign applications will be eventually used for more dangerous applications such as advanced genetic engineering and the military. 3. AI uses too much energy. It's disrespectful to the resources that we have. 4. AI is an apex technology amongst…
All of the above. It is even worse than crypto. The AI proponents that support this have no serious solution to the mass displacement of jobs thanks to AI. They actually don't mention any alternative solutions and instead scream about nonsense such as UBI which has never worked at a large sustainable scale. > Society needs to slow down and find alternative, more sustainable solutions. AI is aligned with short-term ec…
Earlier quoted context omitted.
Jetbrains is still the leader in all of the small details that make navigating and working a code base easy. Copy and paste workflow is a minor slowdown, but nothing compared to things like smart links in terminal, auto detection of run configurations, etc, etc.
Can you give some examples? I spend the majority of my time reading code nowadays, and I often have like 8 or more Sublime Text windows open (each open on a separate codebase). I cant imagine how much RAM it would take to do that in CLion or Visual Studio. Sublime Text's text search is the killer feature for me. CTRL+SHIFT+F and I can search a million LOC+ codebase instantly, navigate the results with CTRL+R, do a su…
Notice that if you work with large projects it is crucial to give the IDE enough RAM so it doesn't thrash a lot. You can also remove a lot of its default plugins to make it much faster.
I ditched Copilot for Cursor and never looked back. Cursor might be the only other AI product with product market fit than ChatGPT, it's that good
Has anyone who uses an IDE (e.g., JetBrains, not a code editor) moved to Cursor? I've downloaded it a few times because everyone raves about it, but I've always come back almost immediately because editors can't reliably make changes across projects (among many other things)... What am I missing? FWIW I use GoLand w/ Supermaven, currently.
Personally, I'm using Goland for editing, along with Zed for its AI assistant. I just have the same project open in both, and do any AI editing in Zed.
I really like the AI UX in Zed with the chat window on the right being context for inline assist (and being able to easily include tons of files in the context window).
Earlier quoted context omitted.
A software developer's time is much more precious than wasting time on sub-optimal models. Open Weights models has it's place (in training custom agents and custom services), but if you are knowledge worker, using a model even 5% less than SOTA is extremely dumb
100% disagree with this take, the flexibility in controlling the prompt leads to QwenCoder2.5-32b outperforming gpt-o1 and claude sonnet 3.5 for nearly everything that I use it for (true for Gemma-27b and llama3.3-70b, though in this context I'm almost always using the former). A specialist model that's specifically prompted to do the correct thing will outperform a SOTA generic model with a one size fits all system…