I saw a video a while back on one social media site or another where someone sitting in a car recorded three young men shotgunning some beers on an apartment balcony. The insinuation being that hanging out was cringe, and that the poster had caught some losers in the act.

It's hard to gauge "real" general sentiment from social media, but if having a beer in a slightly silly way is the level of vulnerability at which you can be recorded for public ridicule, it's not hard to empathize with a generation reluctant to reach out for connection.

We just meant in the city of Cannes.

Are the dividing lines around personality? Working domains? Opinionated software stuff?

Who knows?

I think this is the first time we've had a third minor version bump on a frontier Anthropic model. (I count the 0.5s as major here, because they've been issued non-sequentially and also corresponded to massive capability leaps, eg, Sonnet 3.5, Opus 4.5).

So now the Opus 4.5 family has successors 4.6, 4.7, and 4.8, each posting fairly modest claimed gains. My own experience w/ 4.6 and 4.7 are that I don't firmly grasp any capabilities improvements over my memory of 4.5, but it's all so fuzzy that it's truly difficult to tell.

Maybe my own tastes are saturated now (it's smarter than me?) and I'll never again perceive model progress. Maybe the incrementalism is such that I'd notice immediately if my 4.7 workflows were redirected now to 4.5.

Difficult spot for the labs to be in because, if they have a stronger product, I'd prefer they release it and that I can use it.

But as this dynamic continues, the improvements are going to be less and less legible for end-users, who will complain about the churn-without-payoff, even when the payoff may actually be real.

At a glance, they're providing an interface to YT sourced content with some value adds around tracking or categorizing listening.

A quick question for users: can the site itself be configured as a listener without streaming / displaying the video? In general, YT has a lot of music, but the perf hit of streaming typically high-quality video as well is a blocker when doing dev work on my main machine.

For a while I expected there could be a good return on a good implementation of this, but now as soon as a strong interface itself is created it seems easy to copy.

On balance, and via dictation, it feels likely to be faster overall to just enact the changes I want 'inline' of the conversation thread.

Is this stuff any better now? I think current harnesses probably do have things like file change listeners that automatically inform agents before they act on a file they've previously engaged with if it has changed in the meantime.

On average, I wouldn't expect the regulatory agencies to be very friendly toward outright fraudulent reporting from Waymo. On the very outside, maybe these 80-90% reductions are optimistic roundups from 50-65% reductions. Or do you believe that Waymo is secretly running people down and scooping corpses into their trunks?

What is a sedentary pace of driving?

I don't intend this to read as pure snark, but someone's abstract value isn't much good to them if the job market itself can't / won't recognize it.

We currently sit in the ballpark of 300,000 pedestrian deaths per year worldwide [2]. You should be relieved every time they deploy to a new city.

[1] - https://waymo.com/safety/impact/

[2] - https://ourworldindata.org/data-insights/more-than-a-million...

https://waymo.com/safety/impact/

Over a given driving distance, compared to humans, Waymos produce a 90% reduction in serious injury, 90% reduction in pedestrian strikes, 83% reduction in airbag deployments, 85% reduction in cyclist strikes.

The lack of compassion that people display here is shocking to me.

"Don't automate science, because there are junior scientists could be denied the thrill of specific discoveries."

Cancer patients are not accessories to anyone's self-actualization.

Having models attempt an SVG letter S remains one of my personal/informal LLM benchmarks. They are still pretty bad at it.

EG, my own oldest child needed a surgery at birth that would have been logistically impossible even 50 years ago. I'd say that she and I have benefited enormously, despite not being billionaires.

edit: I solemnly swear that the sibling comment with the strikingly similar "impossible 50 years ago" claim is a pure coincidence and that I at least am not a bot campaign. Haha.

Even with LLMs, posts like this don't just fall out of a coconut tree. If you have a set of target benchmarks for your own model, then keeping "the set" of side-by-side comparable models is its own maintenance headache.

Pol here is abbreviated politician.

Practically nobody teaching K-12 has subject-matter masters degrees. It's just not part of the career trajectory. As unusual as a nurse having an M.A. in history or something. Yes, would occur on the margins of people changing course in life, but not the mainline.

Specifically, the question here is about the efficacy of pay-scale bumps for Masters degrees in education. To your point (and my counter-point), teachers get a substantial pay bump* if they hold a M.Ed, but no bump if they hold a masters in their teachable areas.

For persons who can afford it in the moment, taking a one year or two or three year part-time M.Ed. after getting a few years teaching experience (an entrance requirement in most M.Ed. programs) can pay for itself over the next 2-5 years, then is all surplus for the rest of the career.

* - all of the varies a bit by jurisdiction but I think this is "the general case".