felarof

10 ms·

121.
by felarof 1y ago
Yeah, I feel LLMs can finally solve the tab overload issue. I suffer from this constantly. I added few features which I felt would be useful - easy way to organise and group tabs - simple way to save and resume sessions with selective conte
122.
by felarof 1y ago
Yes, eventually we think there is more value of owning the entire stack than just be a MCP connector. Few ideas we were thinking of: integrating a small LLM, building MCP store into browser, building a more AI friendly DOM, etc. Even today,
123.
by felarof 1y ago
Yes, it works.
124.
by felarof 1y ago
Hopefully in a month or two. Sorry!
125.
by felarof 1y ago
Qwen3 8B works pretty well. But for complex planning and navigation tasks, big models (GPT4.1, claude 3.7) are the still the best bet. We also let you use your own API keys for the big models.
126.
by felarof 1y ago
Thank you! Yeah all user data is just stored locally on device. Oh cool, will look into basic.tech to understand more.
127.
by felarof 1y ago
Thanks for asking - not a stupid question at all! I should have probably explained it at the top of my post. By "agentic browser" we basically mean a browser with AI agents that can do web navigation tasks for you. So instead of y
128.
by felarof 1y ago
Sounds good! will look into getting linux build.
129.
by felarof 1y ago
Yes MacOS for now, but looking into getting Linux binary next. > what's the reason for no Linux/Windows? Sorry, just lack of time. Also we use Sparkle for distributing updates, which is MacOS only. > Also what's the bus
130.
by felarof 1y ago
> 100% runs locally Thank you! We have ollama integration already, you can run models locally and use that for AI chat.
131.
by felarof 1y ago
Thanks for the feedback. Honestly—we just reused the icon we had gotten professionally designed for the last idea we were working on ( https://felafax.ai/ ). But not gonna lie, as a tiny startup, we don’t have marketing budge
132.
by felarof 1y ago
No, not today. But wonder if it matter if it the agent is mostly using it for "human" use cases and not scrapping?
133.
by felarof 1y ago
16GB M1 Pro is good enough to run our browser! You should give it a try! Download form https://www.nxtscape.ai/ or our github page.
134.
by felarof 1y ago
Haha, used gpt4o to generate it. What change do you want to see in that fox appearance? Any change should be a prompt away :)
135.
by felarof 1y ago
Today, we connect to chrome using CDP and use Puppeteer to send clicks and other operations. Also, using browser use DOM tree highlighting, which works great. To get the page content we parse accessibility tree.
136.
by felarof 1y ago
Thanks for the feedback. I get the general sentiment. But cursor for sure has improved productivity by a huge multiplicative factor, especially for simpler stuff (like building chrome extension).
137.
by felarof 1y ago
Yes, we are building on top of chromium. haha noway two of us can build a new browser and feel it's not needed too.
138.
by felarof 1y ago
There is a big red button to always stop the agent.
139.
by felarof 1y ago
Haha, was easier to build and we were the first users :) have linux next on our radar. What build do you want?
140.
by felarof 1y ago
Yeah accessibility is one such usecase, but in future we have few other ideaswhere having a fork makes it lot easier. Few ideas: - Ship a small LLM along with browser - MCP store built in
141.
Show HN: Nxtscape – an open-source agentic browser(github.com)
314 points by felarof 1y ago|206 comments
142.
by felarof 2y ago
> The big question is whether demand for CUDA can be supplanted with application-specific accelerators. At least for AI workloads, Google's XLA compiler and the JAX ML framework have reduced the need for something like CUDA. There a
143.
by felarof 2y ago
> Which means barrier to entry is very very high. +1 on this. The tooling to use TPUs still needs more work. But we are betting on building this tooling and unlocking these ASIC chips ( https://github.com/felafax/fela
144.
by felarof 2y ago
dumb question: wdym by sparse work? Is it embedding lookups? (TPUs have had BarnaCore for efficient embedding lookups since TPU v3)
145.
by felarof 2y ago
+1, almost all (if not all) Google training runs on TPU. They don't use NVIDIA GPUs at all.
146.
by felarof 2y ago
> CUDA, just the subset of tensor operations that are used for training and AI inference. If demand for training and inference wanes Interesting take, but why would demand for training and inference wade? This seems like a very contrari
147.
by felarof 2y ago
TPUs have been used for training since a long time. (PS: we are startup trying to make TPUs more accessible, if you wanna fine-tune Llama3 on TPU check out https://github.com/felafax/felafax )
148.
by felarof 2y ago
+1, we still have a lot of performance we can extract! JIT-compiled train steps, more optimized data loading and sharding, gradient accumulation, and activation checkpointing. We will continue building and will do another blog soon after im
149.
by felarof 2y ago
Tinygrad is great, but still in early stages I believe. JAX has matured a lot over last 6 years and XLA has been around for lot longer. We believe we can extract good perf from AMD with JAX + XLA kernels.
150.
by felarof 2y ago
We have a few technical issues that we still need to address: 1) This entire fine-tuning run was done in JAX eager mode. I kept running out of memory (OOM) when trying to `jax.jit` the entire training step. Even gradual `jax.jit` didn'

More ›