pzo

17 ms·

301.
by pzo 1y ago
I think it's simplification to compare progress only on LLM level. We had big progress in AI in last 2 years but have to take into account more than text token generation. We have image generation that is not only super realistic but y
302.
by pzo 1y ago
There is also Trae - vscode fork from bytedance
303.
by pzo 1y ago
also for some reason I tested simple prompt (few words, no system prompt) with attached 1 images and openrouter charged me like ~1700 tokens when on the other hand using directly via python-genai its like ~400 tokens. Also keep in mind the
304.
by pzo 1y ago
but if you want to use google SDK (python-genai, js-genai) rather than openai SDK (If found google api more feature rich when using different modality like audio/images/video) you cannot use openrouter. Also not sure if you are de
305.
by pzo 1y ago
"The previous iteration (03-25) now points to the most recent version (05-06), so no action is required to use the improved model"
306.
by pzo 1y ago
I feel your pain. Cursor has docs features but many times when I pointed to check @docs and selected one recently indexed one it sometimes still didn't get it. I still have to try contex7 mcp which looks promising: https://g
307.
by pzo 1y ago
This looks great will definitely have a look. I'm just wondering if you tested fastRTC from hugging face? I haven't done that curious about speed between this vs fastrtc vs pipecat.
308.
by pzo 1y ago
This looks great will definitely have a look. What I wish that there was also MCP client (SDK) for mobile devices either native (Swift / Kotlin ) or React Native / Expo ( ideally with integration with vercel ai sdk). For mobile de
309.
by pzo 1y ago
I think OP means that FOSS didn't digest many multimodals of phi4-mini-multimodal such as Audio Input (STT) and Audio Output (TTS), also Image Input also not much supported in many FOSS.
310.
by pzo 1y ago
yeah especially that this simplifies e.g. doing mobile app for 3rd party developers - not extra cost, no need to setup proxy server, monitoring usage to detect abuse, don't need to make complicated subscription plan per usage. We just
311.
by pzo 1y ago
With current geopolitical environment I would be more inclined to switch to Kagi if they move HQ to EU or Swiss.
312.
by pzo 1y ago
What kind of services? IM app? Thailand uses LINE, Vietnam: Zalo, Cambodia and Myanmar Telegram and Viber, in Indonesia many uses few IM at the same time.
313.
by pzo 1y ago
Kokoro gives great results especially when speaking english. Model is small enough to run even on smartphone ~3x faster than realtime.
314.
by pzo 1y ago
Sounds great. Hope more language support in the future. In comparison Sesame CSM-1B sounds like trained on stoned people.
315.
by pzo 1y ago
windsurf also have plugins to jetbrains - they rebranded the whole company from codeium to windsurf and their jetbrains plugin also support cascade.
316.
by pzo 1y ago
OpenAI shared chart about performance drop with large context like 500k tokens etc. So you still want to limit the context not only for the cost but performance as well. You also probably want to limit context to speedup inference and get r
317.
by pzo 1y ago
there was news they recently bought a lot of nvidia gpus since their progress was too slow to use their own chips even in their own data centers for their own purposes
318.
by pzo 1y ago
This is still useful if you combine this method with other methods to make depth map more dense and metric e.g. photogrammetry or ML depth estimation model. AFAIK this is how apple depth api works with their lidar. Here bytedance release ve
319.
by pzo 1y ago
I had the same exact experience - after disappointment (couldn't use like 2/3 of my premium credits because every second request failed after they upgraded to 0.46) unsubscribed. They offered refund in email. I replied I wanted re
320.
by pzo 2y ago
Did you benchmarked combo: DeepSeek R1 + DeepSeek V3 (0324)? There is combo on 3rd place : DeepSeek R1 + claude-3-5-sonnet-20241022 and also V3 new beating claude 3.5 so in theory R1 + V3 should be even on 2nd place. Just curious if that w
321.
by pzo 2y ago
they share in livestream that 4.1-nano is worse than 4o-mini - so nano is cheaper, faster and have bigger context but worse in intelligence. 4.1mini is smarter but there is price increase.
322.
by pzo 2y ago
yeah and considering that gemini 2.0 flash is much better than 4o-mini. On top of that gemini have also audio input as modality and realtime API for both audio input and output + web search grounding + free tier.
323.
by pzo 2y ago
yeah I was surprised in they benchmarks during livestream they didn't compare to ChatGPT-4o (2025-03-26) but only older one.
324.
by pzo 2y ago
If you are entering a gas station you are only for ~5 min. In a cab you are sometimes for half an hour - If somebody call you during a ride it's weird to tell "sorry I will call you later because I'm in a cab and being record
325.
by pzo 2y ago
yes they also have very decent free tier which is great, but keep in mind prompts "used to improve our products". We will see if free tier will stick for long or is temporary - they removed details how big free tier is but still g
326.
by pzo 2y ago
> It might be relevant that I've completely stopped using Cursor in favor of other tools/agents. Can you share what you use these days? I switched from cursor to windsurf but also want to play more with Trae and Cline/RooC
327.
by pzo 2y ago
Apart from Gemini 2.5 Pro they have a decent Jack-of-all-trades master of none/price Gemini 2.0 Flash. 1) is dirty cheap ($0.1M/$0.4M), 2) is multimodal (image and audio), 3) reliable rate limit (comparing to OSS ml ai providers),
328.
by pzo 2y ago
> WASM-powered onnx runtime background removal tool Yes it's slow because they will most likely execute on CPU (for sure on iOS). WebGPU is still not enabled on safari and WebNN is not even supported anywhere (so that you can use NP
329.
by pzo 2y ago
few old iphones or android smarthphones would do the trick even better - they have much better cameras than webcam, old iphones can even record at 240fps. NPU, GPU, LTE, GPS, battery, IMU, stereo microphones also already included in one pac
330.
by pzo 2y ago
"models are released under Apple ML Research Model Terms of Use."

More ›