Linas Beliūnas’ Post

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.

This is huge! French AI lab Kyutai just released Moshi, an open-source GPT-4o competitor that can listen, hear, and speak 😳 Moshi is the first real-time multimodal LLM that operates with basically no delay (end-to-end latency is 200ms). In fact, this Voice AI even interrupted the speaker a few times 😅 The best thing about this is that, unlike OpenAI's GPT-4o demoed in May, Moshi will be open source with code, model, and paper to be released soon. But here comes the crazy part. Kyutai built this end-to-end voice & text model in just 4 months with a tiny team of 8 people. French engineers are literally insane 🤯 Soon this Voice AI tech will be everywhere and change the way we do things forever. P.S. for more interesting stuff, check out 🔔linas.substack.com🔔, it's the only newsletter you need for all things when Finance meets Technology. For founders, builders, and leaders.