Simli retweeted
Congratulations to @simli_ai on the release of Trinity-1, an interactive AI avatar API with <300ms latency ⚡️
Learn how Simli’s team increased avatar sessions per $ by 2-3 times and reduced GPU startup times from 5 minutes to under 2 with DataCrunch ⬇️
datacrunch.io/blog/how-simli…
Simli retweeted
🚀 Integration with @simli_ai is LIVE!
@simli_ai x VideoSDK. Bring your AI voice agents to life with hyper-realistic avatars. Lifelike. Interactive. Real-time ready.
🔗 Check it out: docs.videosdk.live/ai_agents…
Simli retweeted
Nice videos from @imkylecampbell walking you through two building two fun @pipecat_ai agents.
➡️ A Voice AI phone agent with @twilio
➡️ An interactive video avatar with @simli_ai
Simli retweeted
Whoa, the CEO of @trydaily just shouted my videos out! Thanks @kwindla! Love the tech behind @pipecat_ai agents and glad others are finding my videos helpful.
More to come!
AI Agents = @pipecat_ai + @simli_ai / @twilio
Videos: youtu.be/rQ-S0q8EZmM?si=bZg-…
Nice videos from @imkylecampbell walking you through two building two fun @pipecat_ai agents.
➡️ A Voice AI phone agent with @twilio
➡️ An interactive video avatar with @simli_ai
Simli retweeted
We are fundamentally changing human computer interaction with realtime AI avatars, cool to see @a16z do make this overview with @simli on it.
Creating a believable AI avatar – with a realistic face and voice that combine to make a talking character – is challenging.
But it’s happening.
We’re beginning to see lifelike AI characters and human clones in advertising, content creation, and corporate communications.
@venturetwins on what’s working now, what’s next, and the most impressive AI avatar products today, after her own hands-on testing: bit.ly/425MgKv
Simli retweeted
Arabic Poetry + @Simli_ai + @elevenlabs
ElevenLabs newest conversational AI agent is pretty cool!
Simli retweeted
Huge turnout for Aimadlab meetup on simulating worlds locally with @MistralAI models , ran out pizza 🥹
Simli retweeted
I just made our lowest latency video agent yet.
Moshi from @kyutai_labs (voice 2 voice) + @Simli_ai (avatar) = consistent 1sec latency on M2 Mac
🍡I love mochi ice cream, turns out moshi does too 😉
Simli retweeted
I run an AI-first team.
If we’ve done something 3 times and we think it’ll happen at least 10 more times, we make a GPT.
If the payback period for a new tool is under a week, we buy it for a year.
If the payback period for a new tool is under a month, we buy it the next month.
We invite AI into all of our brainstorms.
We use AI to design new websites.
We leverage AI for custom code.
We don’t say Google it, we say ask AI.
Nearly everything we do—every meeting, every newsletter, every contract—is supported with AI.
(Equally important: we have a policy of when NOT to use AI.)
We spend thousands of dollars a year on AI tools and compute.
And it’s worth every penny.
Simli retweeted
I asked Einstein about doomer vs e/accs. Einstein seems to be in the Yann LeCun camp when it comes to AI and p(doom).
This was made with, @DeepgramAI STT, Llama8b on @GroqInc , @elevenlabs TTS and @WeAreSimli for the avatar api.
Simli retweeted
Memes are fun - which meme would you want to talk to?
Our speech to talking avatar API is down to 250 ms ttfb. That makes Reat-time video interactions possible. Its in Alpha, give it a try.
Entire unoptimized stack consists of, deepgram stt, gpt-4o-mini, elevenlabs and simli.
Simli retweeted
Job interviews can suck.
We put together a quick and dirty mock job interview AI voice/video bot, to help you practice.
@GenAICollective and @trydaily are putting on a show with the Most interesting Bot in the World Contest. Super psyched for the event tonight!
The tech stack: azure stt, gpt-3.5 / 4, elevenlabs tts, simli for the video, and daily for webrtc magic.
What would make this better?
You can try it at simli.com/job-interview
@WeAreSimli















