Introducing Halo, the best framework for post-training of open-source models.
Halo delivers up to 2.8x the throughput of stock TRL with less peak memory, while models stay in their native HuggingFace format.
Star us on GitHub: https://t.co/3mAiUljdrN
I appreciate this is a popular narrative, but my brilliant historian friend argues it’s a fundamental misread of what happened. In fact the real interpretation is even more perilous.
Why not reform urgently?
They did, they went bankrupt in attempting to do so urgently
Why not slash the bureaucracy?
Austrians hyper consolidated and streamlined within cisleithania in a mass modernisation effort, and transformed a place like bosnia in 20 years. The ottomans had no real bureaucracy to speak of when the decline started and had to speed rush the civil service to establish one to survive.
Race to adopt new technology?
Yeah the austrian military tech and industrial base were outrageously advanced at the start of the war due to literal years of the state putting everything it had on the line. the ottoman state’s official newspaper was literally called the “bounties of science” and every single school they opened was focused on the hard sciences after the 1860s, and they sent tens of thousands to learn western science and the apply it. The ottoman sultans personally ordered the destruction of 50% of the topkapı palace to build a railroad.
Austria and the Ottomans were doing everything you possibly could to end their decline, including sacrificing their own culture and heritage to do so. They had strong leaders desperate to turn things around, and it was all in vain. Solving these issues is extremely difficult, which makes our current regulatory and government incompetence all the more dangerous.
As a teenager reading about the decline of the Ottoman and Habsburg empires, I couldn't understand why they mostly just let it happen. Why not reform urgently, slash the bureaucracy, race to adopt the new technologies?
Much less confused these days.
Actor has raised a $4M seed round. I am excited to show a snapshot of what our deployed robot models can do in the real-world.
We're incredibly grateful to be supported by investors from Eniac Ventures, Hyperion Capital, Hummingbird VC, 2048 Ventures, Nova Fund, Vanderbilt Alumni, and many incredible angels.
If you are interested in deploying machines with us, contact me : )
Easier to see in the “startup”/founder stage, but Lip-Bu Tan is a fantastic example of Great Man theory playing out and breathing new life into old dinosaurs. Sure there are macro tailwinds, but that hasn’t helped Qualcomm, Broadcom, or ARM nearly as much.
Once asked an early 🇨🇳 robotics founder if he built his own motors. Turned out his team had even created a custom winding machine because the specs wouldn't have been good enough otherwise. Would love to see more of this "no excuses" attitude in 🇺🇸 + 🇪🇺.
@SeanZCai Ability to differentiate in the far right tail normal distribution of signals is all that's important when the success drawn from that distribution is a power law. Certainly true in VC, probably also true in RL.