Great summary of the important distinction between distillation and training on the output of other models.

I argued that chemical reaction speeds etc. will be a real blocker in my anti-doomer screed here: addxorrol.blogspot.com/2024/…

every time opencode gets posted on HN there's all these idiots who complain about me being mean to anthropic like i have some extremist viewpoint ben is extremely reasonable, very smart and this article understands the ecosystem better most of what i read and he's concerned

Tim Hortons LLM is currently being trained on a cluster of 200,000 unsold blackberry Z10s

The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…

We have hockey for fighting.

If we meet aliens will we assume they are conscious?

It’s genuinely curious how they arrive at this new vocabulary? It reminds me when I moved to Australia where they have “smash repairs” to fix automobiles. I knew what it meant but it was strange to my Canadian sensibilities. They are a foreign country.

Seems reasonable to hold Google liable.

Often want this and hate having to always do it from first principles.

I want to say a final thing about my Fable first reaction: I dedicated my life to programming and I'll use every innovation in the field, also to extract value and bring it to the local inference world, to Redis, and so forth. But:

One of my personal favorite features announced at WWDC will I suspect be a sleeper hit: container machines, allowing your Mac to run a lightweight, persistent Linux environment with your home directory and repos automatically mounted: github.com/apple/container/b…

Is Anthropic succeeding just because they are better at naming things?

I used the new AWS MCP server this week, and it’s MUCH better than earlier iterations. Mostly because it has code mode! I used it to find and nuke unused resources in our dev account. And to find cost optimizations in prod account. aws.amazon.com/products/deve…

“This wasn't hard. But it was nuanced.”

Automate the manual QA steps.

Transformers don’t really have a scan mode do they? Like they literally have to fully process every token? Context: reading a utility bill I know to skip all the boring detail unless I need it for some reason.

I was recently diagnosed with a rare autoimmune disorder called anti-NMDA receptor encephalitis. It's a "disease of chaos" that completely upturned my life for a couple months. I wrote a blog about it that goes into more detail and discusses prognosis. burntsushi.net/encephalitis/

It’s always an alignment problem