have been using @impeccable_ai for the past few days - game changer for thinking through designs and applying my own taste vs. generic AI templates. real design vocab for coding agents, and the workflow is slick.
quick snips of modules from "Iron Buddy," an app I've been tinkering on
I’m excited to announce @covaldev raised a $28M Series A, led by @NorwestVP with participation from @Base10Partners, @twilio Ventures and @ycombinator, among others.
I started Coval with the belief that the simulation and evaluation approaches that made autonomous vehicles a reality could unlock the same for voice AI. That parallel has only gotten sharper as enterprises deploy autonomous conversational agents at scale.
Voice and chat are the new interface in the AI era. Talking is the most natural way to interact with complex autonomous systems. The same way web and mobile transformed every enterprise, every company will have a voice agent as the front door to their product. This platform shift is happening now.
From Y Combinator to working with companies like @DeepgramAI, @perplexity_ai , and @Zoom, it's been a wild ride. We watched voice AI go from an emerging modality to every Fortune 500 enterprise building a conversational interface.
Autonomous agents are the future. But you can't ship and hope for the best when agents are taking on real work. The stakes are too high.
Coval will be what enables us to actually trust those agents. The platform to help teams reliably scale voice and chat agents.
We wouldn’t be here without Coval’s founding engineers @rheacoval and Kobi Hudson. Rhea took a chance on Coval before we had a single customer, a line of code, or an office that wasn't my couch. Kobi left @Waymo after 10 years building foundational simulation systems to help us build the frontier of voice simulation.
To the full Coval team - you all are insane engineers and builders who inspire me every day. Rob Young Alejandra Vergara Henry Finkelstein Callum Reid Dakota Mallen Dana Dzik Mallory McLoughlin Christopher Kuester Jake Levi Brooke Hartley Kappi Patterson
To all our investors - @NorwestVP, @Base10Partners Partners, @twilio Ventures, @ycombinator, @alumniventures, Swift Ventures, Fortitude VC and so many more. You are what made Coval possible.
To @koh_terai who brought my video vision to life.
And last but certainly not least, thank you to everyone who's been building with us.
We're just getting started.
In self-fulfilling prophecies, we start by believing in something false. Our propaganda and self-hypnosis changes our behaviors, and those behaviors then turn something false into something true.
Furthermore, the Pygmalion effect states that high expectations lead to improved performance, while the Golem effect states that low expectations lead to decreased performance. Our expectations of ourselves affect our performance, and the expectations of those around us also affect our performance.
The takeaway is we should focus on positive mindsets, uplifting communities, greater expectations, and the Pygmalion effect. We can't do anything just because someone expects us to, but high expectations from a supportive community can help us achieve more, especially if we are confident that our stretch goals are achievable.
loved this breakdown of Mito in action - amazing what you can build with just a narrative and model stack. always inspiring to see what @inakib + the @mito__ai team are building!
Every frame in this spot started as an intention.
A grip. A rim. An empty gym. A return.
In our latest Behind the Screen, we break down the full creative and technical process behind "The most important part of the season...is the preseason", shot by shot, model by model:
> 10 shots. A single narrative arc: origin, sacrifice, return.
> Multi-model stack: Kling 3.0 for cinematic physics, Seedance 2.0 for jump rope motion, Nano Banana 2 for statics.
>One enforced visual language: single lateral light source, 35mm Kodak grain, radical solitude in every frame.
No crew. No location scout. No casting call. Just a narrative structure and a clear visual language, prompt by prompt.
If you have a story that lives in your head but nowhere else, MITO AI is where you build it.
The preseason is now.
Full breakdown in the first comment 👇
@SahilBloom life isn't fair. you can do everything right and still come up short.
don't let that deter you. keep working hard, cherish the people you love, and help others along the way.
keep hearing the same pattern: junior FTEs are AI-maxxing to increase leverage, while seniors are under pressure to do more with less. results in a growing focus on efficiency everywhere - hiring plans, SW spend, AI infra, etc.
cool to see what teams are building in the inside + how @tryramp , @OpenRouter, and others are benefiting form this shift.
Super insightful Baseten analysis!
LoRA recovers ~98% of full fine-tuning quality training just 3-13% of params >>> as customization gets cheaper, the moat isn't just fine-tuning, but the proprietary eval + data loop you train on.
1/ We fine-tune a lot of customer models, so we decided to systematically try and figure out some best practices for finetuning. SFT isn't sexy, but it's still important. We vary one SFT lever at a time across 2 model families, dense + MoE to 235B, on 4 real-world customer datasets.
What makes this clean is that each dataset is paired with an eval that took weeks to build with the customer, and the training outputs were generated to pass that eval. So the supervised target and the thing we measure downstream are the same criterion, which strips out the usual confounders