It could be the case that we’ve reached the last generation of frontier models that can be accessed by the general public. That eliminates a risk that Anthropic could be leapfrogged by a competitor.
Now it’s a competition between products on the near frontier. Anthropic has executed well on products so far. They blew up thanks to Claude Code, not Opus by itself.
I lean libertarian but I can recognize the danger in having access to a machine that can craft pathogens to spec.
A pathogen with a very long incubation time and a high fatality rate would be about as bad as nuclear war. Maybe we need to figure out how to possibly defend against one person doing this before making it easy for anyone to do it.
I realized that should I end up getting laid off soon I won't have an unlimited token budget and for the workflows I've settled into it would be quite expensive. So I was exploring what it would take to run open models at home.
Was quite disappointed to see that the PC side hasn't kept up. The unified architecture on Macs makes it very hard to justify spending money on a Linux machine for inference workloads.
I'm enjoying it. It's wild to realize that I spent countless hours playing Theme Park when I was around 10 years old, and Demis had been a big contributor to the game when he wasn't much older.
Also I don't really care that it's a bit of a cheerleader for DeepMind and Hassabis. Substantive criticism is good, but too often with these kind of books it feels like an editor told the author that the book needs something negative and the author has to inflate an issue to meet the requirement.
The first time we traveled domestically in China I kept thinking that my wife had to be mistaken, there has to be some kind of confirmation we need to show in order to board. But nope, it literally is just show up with your ID.
Risk tolerance is a value judgment, not an empirical fact waiting to be discovered.
Competent experts could tell you how much safer you would be if you wore a helmet to drive your car. They can't tell you how much you should value that extra bit of safety.
They've changed the AVCOAT to be less permeable and altered the re-entry profile.
One of the findings of Artemis I is that lack of permeability led to trapped gas pockets which expanded and blew out pieces of heat shield. The reason for the change to be less permeable is to make it easier to perform ultrasonic testing, not to improve performance.
> Another chart which the Artemis Tiger Team did not intend to show on Jan. 8th, was the figure showing the spallation events as a function of time during the skip entry heating profiles (Figure 6.0-4 of NESC Report TI-23-0189 Vol. 1). In this figure, it was quite clear that the Program narrative they were feeding to the press, that it was the dwell time during the skip which allowed the gases generated to build up and cause the delta pressures which caused most of the spallation was, again, patently false. In fact, during the first heat pulse (t ≈ 0 to 240 sec), approximately 40-45% of all the medium to large chunks of ablator spalled off the Artemis I heatshield.
> Hence, varying the trajectory would do little to prevent spallation during Artemis II. I was never shown the new, modified trajectory at the Jan. 8th meeting.
It is a bit chilling to watch these astronaut profiles having just read yesterday about the heat shield issues observed on the prior mission, and that this will be the first time we can test the heat shield in the actual pressures and temperatures that it will have to endure.