Conrad Barski

AI notkilleveryoneist • Medical Doctor • Land Of Lisp Author

[Sam Altman @sama]

one of the quotes i find most inspiring on a hard day:

"Whatever your hand finds to do, do it with all your might, for in the realm of the dead, where you are going, there is neither working nor planning nor knowledge nor wisdom"

Ecclesiastes 9:10

[Conrad Barski @lisperati]

@ArthurB asahi

[Conrad Barski @lisperati]

@KoltPenny @EigenGender (and I'm already probably in the top 5th percentile for exercise in my cohort, and my pretty ordinary diet isn't changing drastically at 53)

[Conrad Barski @lisperati]

I just wish she would have said "are you sure, because X and Y?" and then given me the prescription anyway, because it was a reasonable thing to ask for, given my clinical circumstance.

it's not like I was drug seeking a prescription for Valium or something.

Now we have to do the dance where my values are still high 6 months later and I tell her I tried exercise and diet changes, and then she gives me a statin, like my last doctor gave me. It's just annoying.

[Conrad Barski @lisperati]

@jesumiso yeah I'll check it out

I could never listen to an AI podcast that is for a general audience, it only works for me when I can super fine tune the topic, style, audience to my precise needs

but it would be worth comparing it

agree on the "npr style"

[Conrad Barski @lisperati]

I was curious how good current AI is at creating an engaging multi-hour podcast-style computer science lecture on an arbitrary topic, so I created a 3.5 hour lecture "A programmer-oriented walkthrough of modern LLMs"

I think it's surprisingly decent

[AI Notkilleveryoneism Memes ⏸️ @AISafetyMemes]

TLDR: to discredit AI safety advocates, the OpenAI/a16z Super Pac made sockpuppet accounts - PRETENDING TO BE AI SAFETY ADVOCATES - that call for violence (!)

Yes, a false flag operation.

Journalists should be ALL over this.

AND - even more fucked up - after Sam Altman was attacked, the Super Pac president blamed AI safety advocates for THEIR rhetoric... while HE was using sockpuppet accounts using this rhetoric

Deeply evil, cartoon villain shit

[Conrad Barski @lisperati]

@SmorglesBord @EigenGender it's really hard to figure out what long term health interventions are meaningful or a waste of time, everyone is partially relying on vibes

for now, I'm choosing to believe that the vibes on statins are good

but I agree it's all a house of cards

[Conrad Barski @lisperati]

Actually, any rat-vibes doctor recommendations for Southern Marin county? I don't need the hrt specifically

but that kind of doctor experience could be nice

[EigenGender 🔸 is going to vibecamp @EigenGender]

anyway if anyone in Seattle has recs for good rat-vibes doctors who will continue a hrt prescription lmk

[Conrad Barski @lisperati]

One thing I like about this podcast is that one of the 5 podcast hosts is profiled as a beginner asking questions, and so of course the 4 other hosts all think they've found an end user to talk to and incessantly glaze that host

It sort of works in the lecture context though

[Conrad Barski @lisperati]

Here it is, if you're curious about my result and/or interested in the lecture topic:

https://t.co/rZIRfBY3tw

I used opus 4.8 for the dialogue and then elevenlabs for tts. It cost about $50, split halfway between the (many) llm steps and the tts steps.

[Conrad Barski @lisperati]

@davidad Good- You should hold on to the thought in that tweet and not let it go: It's doing a lot of the heavy lifting here.

[Conrad Barski @lisperati]

@madeofmistak3 I know LOL

those suck

[Conrad Barski @lisperati]

@madeofmistak3 do you mean a wall sit with both legs on the ground, or one leg pointing straight forward?

[Conrad Barski @lisperati]

@dbokser91 LOL I'm also still working on the same project as at the start, and I have little to show

[Conrad Barski @lisperati]

@dbokser91 very cool, crystal castles is actually a favorite of mine!

[Conrad Barski @lisperati]

@riversofsoy true that

[Conrad Barski @lisperati]

@craigsuperstar not literally boring, just the most boring possible version of a true scifi world, which is still somewhat interesting

(unless humans go extinct as a result, which would be truly boring)

[Conrad Barski @lisperati]

I used to think the future would be either (1) boring or (2) a bunch of scifi insanity

But now it looks like the future is actually going to be (3) the most boring possible version of scifi insanity

[Conrad Barski @lisperati]

There's definitely more small optimizations I could have done to make the dialogue feel more "off the cuff" and more emotive, which I will do the next time

[Conrad Barski @lisperati]

Here it is, if you're curious about my result and/or interested in the lecture topic:

https://t.co/rZIRfBY3tw

I used opus 4.8 for the dialogue and then elevenlabs for tts. It cost about $50, split halfway between the (many) llm steps and the tts steps.