Yafet Melake

Co-founder of @ExpanseCompute

[Aoden Teo @AodenTeoMT]

Today, we’re excited to introduce Miso One, the most emotive voice model in the world.

Miso One is an 8-billion-parameter text-to-speech model for highly expressive speech generation. It emotes like a human and responds faster than a human, with just 110 milliseconds of latency.

We’ve open-sourced the model weights, with API access coming soon.

Hear how Miso One sounds in the thread below.

[Ismaeel Bashir @ismaeel_bashir_]

1/

Auto-research is taking off. Inside xAI, Anthropic, OpenAI, DeepMind, Meta SI: frontier models are running their own training experiments.

At Expanse, we benchmarked 8 frontier models to see how efficient they are at resource efficiency.

TLDR: They all ended up over-provisioning compute by 2x, even for coding specific models.

[Y Combinator @ycombinator]

Banks can't tell which human is actually operating an account. @Incandor_ai links behavior across your user base, making fraud rings, mule handoffs, and banned operators visible.

Congrats on the launch, @Matthew_Yekell & @RosenzweigLuc!

https://t.co/Bk7Q3X8N3G

[Y Combinator @ycombinator]

Oddpool (@oddpool_alerts) is the institutional data layer for prediction markets. It captures every trade and order book update across every platform and normalizes them under universal tickers.

Congrats on the launch, @c0delemons & @RiteshMalpani!

https://t.co/02YPDfK0hW

[Yafet Melake @yafet_melake]

@ycombinator @trysaffron @roblukan @kazumachoji @_mjyao really cool product. tackling something that definitely needs an update!

[Yafet Melake @yafet_melake]

@ycombinator @saudaraAI @edwardharyono_ Congrats!

[Y Combinator @ycombinator]

Expanse (@ExpanseCompute) unlocks wasted GPU capacity. Submit jobs with the right resources. Optimize them to run faster. Debug failures in seconds. Cloud + on-prem HPC.

Congrats on the launch, @ismaeel_bashir_, @nkdem_b, @yafet_melake, and @erenzmendi03!

https://t.co/3lFcfFk7gS

[Ismaeel Bashir @ismaeel_bashir_]

Saving $56K/month by compressing checkpoints is cool.

But being smarter about when you checkpoint is even cooler.

Compression optimises the cost per checkpoint. Intelligent triggering eliminates the unnecessary ones entirely.