ipsum2

11 ms·

151.
by ipsum2 1y ago
I recently wrote some Javascript to automate clicking coupons. The website checks for non-human clicks using event.isTrusted. Firefox allowed me to bypass this by rewriting the JS to replace s/isTrusted/true, while Chrome Manifest
152.
by ipsum2 1y ago
This is the second time in less than a year something similar has happened. Previously, a different YC company (Pear AI) copied Continue, changed the licenses, and "launched". https://news.ycombinator.com/item?id=4
153.
by ipsum2 1y ago
That's my point. The ones that left early got a large sum of money. The ones that leave later will get less. That would incentivize people to be the first to leave.
154.
by ipsum2 1y ago
How do you figure? If you assume that Meta gets the state of the art model, revenue is non-existent, unless they start a premium tier or put ads. Even then, its not clear if they will exceed the money spent on inference and training compute
155.
by ipsum2 1y ago
The game theoretic aspect of this is quite interesting. If Meta will make OpenAI's model improvements open source, then the value of every poached employee will be worth significantly less as time goes on. That means it's in the e
156.
by ipsum2 1y ago
This topic is ragebait. I predict the comment section will be pointless name calling than genuine discussion.
157.
by ipsum2 1y ago
Sort of. The key bottlenecks are not in tokenization, but running the actual CUDA kernels. Python actually has very little overhead. (See VLLM, which is primarily in Python). So when people (like deepseek) 'rewrite in C++', they&#
158.
by ipsum2 1y ago
I imagine many people will react only to the headline and not read the article, but: "Meta tells The Verge that, for now, it’s not training on your unpublished photos with this new feature. “[The Verge’s headline] implies we are curren
159.
by ipsum2 1y ago
I think you did a good job CYA on this, but what people were really looking for was a way to opt-out of Google collecting code, similar to the opt-out process for the IDE is available.
160.
by ipsum2 1y ago
Sort of, check out https://github.com/senstella/parakeet-mlx .
161.
by ipsum2 1y ago
My understanding is that they have not implemented an opt-out feature for Gemini CLI, like they've done for VSCode and Jetbrains.
162.
by ipsum2 1y ago
I've spent a long time with models, gemma-3-27b feels distilled from Gemini 1.5. I think the useful coding abilities really started to emerge with 2.5.
163.
by ipsum2 1y ago
When should we expect to see an update? I assume there'll be meetings with lawyers for the clarification.
164.
by ipsum2 1y ago
In my own experience, 2.5 Pro 03-26 was by far the best LLM model at the time. The newer models are quantized and distilled (I confirmed this with someone who works on the team), and are a significantly worse experience. I prefer OpenAI O3
165.
by ipsum2 1y ago
If you want real-time, it requires a GPU, but can be underpowered. CPU is a little slower but works fine.
166.
by ipsum2 1y ago
I've been using Nvidia's parakeet model, it's been better than Whisper v3 large and smaller. Only supports English. https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2
167.
by ipsum2 1y ago
If you're using Gemini CLI through your personal Google account, then you are using Gemini Code Assist license and need to follow the T&C for that. Very confusing.
168.
by ipsum2 1y ago
This is useful, and directly contradicts the terms and conditions for Gemini CLI (edit: if you use the personal account, then its governed under the Code Assist T&C). I wonder which one is true?
169.
by ipsum2 1y ago
There's nothing wrong with promoting your own projects, but its a little weird that you don't disclose that you're the creator.
170.
by ipsum2 1y ago
Sorry, that's not correct. Did you check out the link? It doesn't describe the CLI, only the IDE. "You can find the Gemini Code Assist for individuals privacy notice and settings in two ways: - VS Code - IntelliJ "
171.
by ipsum2 1y ago
If you use this, all of your code data will be sent to Google. From their terms: https://developers.google.com/gemini-code-assist/resources/p... When you use Gemini Code Assist for individuals, Google collects you
172.
by ipsum2 1y ago
From the article: "Some consumer advocacy groups argue the dyes aren’t worth the potential risk because they lack nutritional value." It's the precautionary principle in action. Is it scientific? Not really. But does it make
173.
by ipsum2 1y ago
Everything thats in the blog post is basically well known already. Google publishes papers and gives talks about their TPUs. Many details are lacking though, and require some assumptions/best guesses. Jax and XLA are (partially) open s
174.
by ipsum2 1y ago
Using optical flow for point tracking is obvious, not that the goal of tracking pixels is obvious. Regarding your actual question, there's many use cases. - Tracking players or balls in sports - Surveillance
175.
by ipsum2 1y ago
You can get them non-tinted, probably many people do.
176.
by ipsum2 1y ago
Considering that people use screens that are 360-500fps for noticeable improvements in video games, people can definitely perceive that. Next time you see a high refresh screen, move the cursor around rapidly. It's very easy to tell.
177.
by ipsum2 1y ago
They're AI generated and look quite nice, no obvious artifacts (messed up text). Not sure what the use case is though.
178.
by ipsum2 1y ago
Rob Fergus is one of the creators of FAIR. It makes sense for him to lead it.
179.
by ipsum2 1y ago
Incorrect, they work with military as well as AI labs like OpenAI.
180.
by ipsum2 1y ago
In theory, yes, but in practice, I write code on a 4k OLED display and haven't noticed any artifacting.

More ›