AI Developer Offsec Lead I build and test various STT and TTS models and design voice agents
Joined April 2015
- Tweets76
- Following589
- Followers37
- Likes76
Finally got access to live voice @Muse. Very impressive @alexandr_wang. I think this is the new winner for conversation quality compared to Dots, Grok Bot, and Duet (background voice FINALLY enabled)
Duet is slowly climbing up in my rankings and is more capable.
Muse speech is top notch though. speaks in tldr which is amazing! I can't stress enough that Muse's STT is on par with Wispr flow from all my tests. this is very impressive.
I understand what @theo now means about Muse conversation feeling just right. it's hard to put it in words but it mirrors your conversation style like no other agent
ran lfm3-audio 2b through my bench (85+ open models). streamed first audio is fast: ~55ms warm on a 5090, ~42ms on a 3090. quality is mid-pack: utmos 4.165, 27th of 54, and lfm2.5-audio 1.5b still scores higher at 4.383 5uck1ess.github.io/tts-bench…
update: @WisprFlow works in Claude Duet now. flagged the 3rd-party keyboard gap to @gabemulley sunday night and it was in by Tuesday morning. Thank you for the quick turn around. It's incredible that it's possible to have feedback implemented so quickly.
My AI agent could see and write code, but it couldn't hear a thing. earshot reads my video's audio: loudness, timing, stereo, and what each sound is. It found one of my videos was in mono. Free, runs on a laptop. Music: YuE2 #sucklessai
Tym Rabchuk retweeted
hashcat v8.0.0 released!
The largest update in the history of hashcat, with PCFG attacks, hash recipes that crack hashes without a hash-mode, and bcrypt on FPGA boards.
Release notes: github.com/hashcat/hashcat/b…
Download: hashcat.net/hashcat/
All personal agents pros and cons:
Grok Bot: almost perfect, does everything, multiple agents, best desktop experience, BUT live voice just feels mid tier. Winning formula if they keep going and adding features.
Claude Duet: no GitHub integration, no desktop experience, no 3rd-party keyboard, voice dies on lock, good model tho
OpenAI Dots: BEST voice by far, but doing work just feels off, singular, and kind of gimicky vs pure codex.
Muse: free, fun, healthy limits, tone feels good, no live voice (yet?)
Hermes: self-hosted voice (own project), too much maintenance but full control
Ultimate Agent : Grok Bot + Duet model + Dots live voice + Muse cuteness
Hackers were hitting PHP2Shell before PHP shipped the fix. Wired reports AWS logged exploit attempts four days early. Run a PHP site? Update to 8.5.12, 8.4.27, 8.3.36 or 8.2.35. #sucklessai
Tym Rabchuk retweeted
🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!
Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.
🔗 github.com/deepseek-ai/DeepI…
1/4
Instagram Edits only runs on phones. Concat is a free, open-source CapCut replacement for Windows, Mac, Linux and Android. Auto captions, background removal and voice cloning run on your device. No watermark. #sucklessai
Stop paying for Adobe. Someone's rebuilding the whole suite open source in Rust. The photo editor already opens real PSDs with layers and text. Early alpha, but it's moving fast. #sucklessai
Grok Bot has its own email now. Addresses end in @mail.grokbot.com. Ask your Bot or tag @bot to get one, and grab the shortest name you can. #sucklessai
Tym Rabchuk retweeted
🚨LEAKED: Another alternate ending for Elliot Page in The Odyssey just dropped.
🤖 Made with AI
Tym Rabchuk retweeted
Comparing TTS systems is a nightmare with no single benchmark.
tts-bench tracks all known TTS engines with HTML results:
- 15+ systems, 50+ voices benchmarked
- Windows, Mac, Linux covered
- Real-world performance metrics
github.com/5uck1ess/tts-benc…
@centrone_bonnie @mosaic952 after you win a battle, wouldn't you feel superior automatically? #dominance #power