throwuwu

58 ms·

331.
by throwuwu 3y ago
You think we might be better off burning wood and dung instead of solar and nuclear? Carbon free energy requires high technology global civilization. If you make everyone poorer then say goodbye to science and technology, no one cares about
332.
by throwuwu 3y ago
We need all tech on the table including tried and tested nuclear as well as untested space based solar and geoengineering.
333.
by throwuwu 3y ago
And they’ll follow the same curve that India did and China before them and so on but I’ll bet the time scale will be greatly compressed because they benefit from all of our experience and technology and having so many more countries ahead o
334.
by throwuwu 3y ago
I was referring to the original comment about continuing to try even in the face of a terminal diagnosis. So no, it hasn’t stopped us.
335.
by throwuwu 3y ago
No, make everything especially energy as cheap and efficient as possible so that everyone gets richer and healthier and can contribute to solving humanity’s problems.
336.
by throwuwu 3y ago
Population collapse is like a banking collapse, it happens slowly at first and then all at once.
337.
by throwuwu 3y ago
Until they get smoked by an asteroid again. Even if nonhuman life somehow doesn’t get wiped out by other factors it’ll all die with the sun. If we are the only place where life exists then that means humanity is the only chance life has of
338.
by throwuwu 3y ago
The planet by itself can only support a few million humans. It’s human technology and intelligence that support the rest.
339.
by throwuwu 3y ago
If you think that population trends are still exponential then you’ve been living under a rock
340.
by throwuwu 3y ago
Hasn’t stopped us yet
341.
by throwuwu 3y ago
If you’re really concerned then change careers to a field that is working on tech to improve the human condition. Short list: 1. Nuclear physics 2. Robotics and automation 3. Space engineering and rocketry 4. Materials science 5. Biotech Ev
342.
by throwuwu 3y ago
I’m honestly hoping AI lets us transcend or at least fully control biology. It sucks being trapped in a bag of meat you have no control over whose only optimized function is to reproduce and then die. No thanks, I’ve got more important shit
343.
by throwuwu 3y ago
A high res webcam and OpenCV?
344.
by throwuwu 3y ago
There are dozens of us!
345.
by throwuwu 3y ago
The Quest has been doing this for years with the guardian setup and now with their option to bring your desk and couch into VR. I think the Pro has more advanced features but I haven’t tried it. You do have to assist it by drawing our bound
346.
by throwuwu 3y ago
Simple protocols would enable this for existing hardware and software
347.
by throwuwu 3y ago
That’s a unique take. Especially for guns in VR which feel incredibly natural with a controller that is basically a pistol grip with a trigger. For people really into Pavlov they even sell kits to mount the controllers together into a rifle
348.
by throwuwu 3y ago
Wait, were you the one controlling where he was going? Like zooming him around the planet? Cause that’s a sure fire way to make anyone nauseous.
349.
by throwuwu 3y ago
> more input still leads to more output. Not when people are involved
350.
by throwuwu 3y ago
Needs less rhetoric and more numbers
351.
by throwuwu 3y ago
Thank you for sharing! I’ve never understood the purpose of this activity. I’m pretty sure no one bothers to read them and even if they do they don’t understand them.
352.
by throwuwu 3y ago
I subscribe and I use GPT-4
353.
by throwuwu 3y ago
That’s pretty easy to avoid if you have them work in an office on company provided machines on a network that blocks access to LLM sites. Not exactly rocket science here.
354.
by throwuwu 3y ago
They just put out a blog post about how they’re doing exactly that by rewarding chain of thought style reasoning.
355.
by throwuwu 3y ago
It might be better for the corpus to be expanded with higher quality text that is produced on demand from contractors. There are also large pools of data yet to be accessed. One giant source would be podcast transcripts but there are many o
356.
by throwuwu 3y ago
I guess we could go back through our histories and resubmit the same prompts a few times. Wouldn’t be a fair comparison since we only have 1 sample from the old version.
357.
by throwuwu 3y ago
I tried comparing the API response against the chat on the same question a few times. There isn’t a huge difference but I’d pick the responses from the API over the ones from chat. Hard to say though, could be RNG.
358.
by throwuwu 3y ago
The moderation endpoint is separate for the API so I’d imagine it’s the same for chat. The model could be the same while they change: Moderation Temperature / top-p System prompt Some other internal system we aren’t aware of
359.
by throwuwu 3y ago
It has and so has the YouTube algo. You can try the verbatim setting with search to see that there has been some monkeying going on but even that doesn’t fix everything. Maybe that’s due to spam sites but it just seems less effective.
360.
by throwuwu 3y ago
The refusals are new. I’ve gotten a few myself recently. I’ve never seen them before. They’re not moderation related and they always include some text about the task being too complex.

More ›