Guaranteed Safe AI paper club meets again this thursday

Event for the paper club: https://calendar.app.google/2a11YNXUFwzHbT3TA

blurb about the paper in last month's newsletter:

... If you’re wondering why you just read all that, here’s the juice: often in GSAI position papers there’ll be some reference to expectations that capture “harm” or “safety”. Preexpectations and postexpectations with respect to particular pairs of programs could be a great way to cash this out, cuz we could look at programs as interventions and simulate RCTs (labeling one program control and one treatment) in our world modeling stack. When it comes to harm and safety, Prop and bool are definitely not rich enough.

my dude, top level post- this does not read like a shortform

Yoshua Bengio is giving a talk online tomorrow https://lu.ma/4ylbvs75

by virtue of their technical chops, also care about their career capital.

I didn't understand this-- "their technical chops impose opportunity cost as they're able to build very safe successful careers if they toe the line" would make sense, or they care about career capital independent of their technical chops would make sense. But here, the relation between technical chops and caring about career capital doesn't come through clear.

did anyone draw up an estimate of how much the proportion of code written by LLMs will increase? or even what the proportion is today

I was thinking the same thing this morning! My main thought was, "this is a trap. ain't no way I'm pressing a big red button especially not so near to petrov day"

GSAI paper club is tomorrow (gcal ticket), summary (by me) and discussion of this paper

Alas, belief is easier than disbelief; we believe instinctively, but disbelief requires a conscious effort.

Yes, but this is one thing that I have felt being mutated as I read the sequences and continued to hang out with you lot (roughly 8 years ago, with some off and on)

By all means. Happy for that