AI Voice Hits Inflection Point With Sesame

In partnership with

Hi everyone,

This is Lore, the newsletter that helps you thrive in the age of AI.  

Thank you for being a part of 25k+ techno-optimists and growing. 🙏

P.S. Watch the latest episode of The Next Wave, this week with Matthew Berman!

Run Your Sales on Autopilot

Increase the output of your sales team without buying more tools or hiring new SDRs. Onboard Agent Frank, Salesforge’s AI SDR, to fully automate prospecting, personalized outreach and booking meetings while your team focuses on closing deals. Get a personalised email from Agent Frank to see his work in action:

Show me the magic

1) Things Are Getting Better

New open-source AI model QwQ-32B from Chinese company Alibaba, is around DeepSeek-R1 level, but 20x smaller. People are already running this model locally on MacBook Pros. Models keep getting not only smarter, but also more efficient and cheaper to run. You can try it out right now on Groq(not to be confused with Grok).

2) AI Voice Hits Inflection Point With Sesame

This week, Sesame—a startup backed by A16Z and Spark Capital—dropped an AI voice demo that’s incredible. I'm not exaggerating: we've reached an inflection point.

Stop everything and try it yourself. You won't fully grasp how transformative this moment is otherwise. ChatGPT’s voice mode is helpful but still robotic (maybe intentionally, I think they didn’t want to freak people out). Grok 3 is wild and funny but you can still tell it’s not human.

Sesame? It’s something entirely new. It’s like being able to talk to Samantha from the movie Her.

For the first 10-20 seconds of chatting with Maya (the female voice), I genuinely felt I was speaking with a woman with an attractive voice —complete with realistic breathing and tonal shifts that mirrored mine. It’s uncanny—almost awkward—how real that connection felt. It made me have some real questions about it because, I could see certain kinds of people getting hooked to this feeling of having an emotional connection with an AI.

Sesame is not perfect yet—push beyond a minute, and context can slip—but the realism is already jaw-dropping. Now, imagine combining this tech with robots launching later this year (2 tweets with examples):

We’re going to have machines interacting naturally with us every day. That’s going to be in some people’s houses this year(probably mostly rich tech CEOs). And will likely be common place within 3-5 years in most people’s houses.

I’ve already had a conversation with my wife about what kind of robots she’s okay with us having. I have a feeling human and AI relationships are about to get really weird.

By the way, Matt Wolfe and I tested Sesame live on our latest episode of The Next Wave, laughed so much we cried as Grok 3 went full Andrew Dice Clay, and unpacked what the robotics revolution really means. Subscribe now to catch it when it drops next week!

AD

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

Sign up now for free and work smarter, not harder.

3) Grok 3 vs Claude 3.7 vs GPT-4.5: Which Update is The Best?

In this week’s episode of The Next Wave, Matt Wolfe and I spoke with our friend Matthew Berman about the latest AI models and what they mean for the future of AI. Check it out!

Also, please subscribe to The Next Wave on Apple, Spotify, and YouTube. 

4) GPT 4.5 Is Better Than You Think

GPT-4.5 landed last week for ChatGPT Pro users (and today for Plus users), but the internet mostly shrugged—too slow, too pricey, not a game-changer.

I disagree.

Having used it extensively, it's the best writing model I've tested yet. GPT-4.5 seems to get nuance and understand things at a deeper level than previous models, with in my experience the closest to it right now being Grok 3.

Benchmarks back this up, with GPT-4.5 holding its own against specialized reasoning models despite being "just" a language model juiced up with more training data. (See chart below.)

Marc Chen from OpenAI recently confirmed this aligns with their expectations: there's no wall in sight for scaling performance through more data.

Now, imagine this: GPT-5 by June, which will probably be GPT 4.5 with o3+ reasoning on top of it, as well as advanced voice mode. Meanwhile, Grok is getting better every day, and constantly is vying with OpenAI’s models for the #1 spot.

We're potentially looking at LLMs getting 5-10x smarter within this year alone.

We’re not just accelerating—we’re warping into a post-AGI future. The ride’s about to get wild, so buckle up.

5) Things I’m Learning From

That’s all for today. Please consider sharing the newsletter with your friends if you think they’d enjoy it. 🙏

Please give me feedback on Twitter. I am still figuring out the best format for this newsletter. Which part did you like? Dislike? Do you have other suggestions? Please let me know—just tweet @NathanLands.

-Nathan