Veo 3 and Kling are some of the most amazing pieces of technology I've ever seen. But there’s still one major piece missing from the puzzle when it comes to AI video: Dynamically expressive voice. You can have the most cinematic visuals imaginable, but if the voice sounds
@aibreakfast
-
AI video lacks dynamic expressive voice despite stunning visuals
By
–
Veo 3 and Kling are some of the most amazing pieces of technology I've ever seen. But there’s still one major piece missing from the puzzle when it comes to AI video: Dynamically expressive voice. You can have the most cinematic visuals imaginable, but if the voice sounds
-
Typical Day in Los Angeles 2025, GTA Graphics Style
By
–
Google Veo 3 prompt:
— AI Breakfast (@AiBreakfast) 9 juin 2025
"Typical day in Los Angeles 2025, GTA graphics" pic.twitter.com/B9Fd352Aa0Google Veo 3 prompt: "Typical day in Los Angeles 2025, GTA graphics"
-
ElevenLabs Pays Voice Actors for Uploading Their Voices, $5M+ Paid
By
–
Many people don’t realize that ElevenLabs will pay you to upload your voice to their platform $5M+ in total payouts to voice actors so far
-
Perplexity integrates SEC data for stock queries, citing original sources.
By
–
Massive win for day traders: Perplexity just got a complete (!) SEC data integration, so queries about stocks are now backed with citations and references to original source documents. Start your one-man hedge fund.
-
AI fully generates a presentation of podcast lessons
By
–
Had an AI Agent watch a @hubermanlab podcast for me and compile all of the key takeaways into a presentation. These slides were 100% generated by AI.
-
AI voice surpasses uncanny valley with one-shot style transfer
By
–
AI voice has pole-vaulted over the uncanny valley.
— AI Breakfast (@AiBreakfast) 4 juin 2025
Bland TTS has figured out one-shot style transfer, so you can clone a voice from one MP3 and tweak its emotion, tone, cadence, all of it. It’s weirdly natural.
It will be used to make voiceovers, build apps, or spin up… pic.twitter.com/ZMnqWL3stnAI voice has pole-vaulted over the uncanny valley. Bland TTS has figured out one-shot style transfer, so you can clone a voice from one MP3 and tweak its emotion, tone, cadence, all of it. It’s weirdly natural. It will be used to make voiceovers, build apps, or spin up
-
10/10: The ideal feel for LLM-powered dev tools
By
–
10/10 this is what every LLM-powered dev tool should feel like https://t.co/MfOIjrdTYb
— AI Breakfast (@AiBreakfast) 4 juin 202510/10 this is what every LLM-powered dev tool should feel like
-
Text-to-speech evolution dissolves interface and identity boundaries
By
–
No one took text-to-speech seriously for years. Even at its best, it was still obviously fake. That’s about to change. In the next few months, the boundary between interface and identity will begin to dissolve. It will reshape culture, commerce, and consciousness. We'll
-
Experts clueless about AI’s future in 10 years
By
–
Not even the best minds in the world have any idea where AI will be in 10 years.