gfodor

18 ms·

271.
by gfodor 4y ago
A recent paper from MS implies transformer LLMs are converging onto AGI. So it’s probably a good idea to abandon these stochastic parrot mental models.
272.
by gfodor 4y ago
You're wrong. I wonder if you've tried GPT-4 yet.
273.
by gfodor 4y ago
it should have been clippy
274.
by gfodor 4y ago
Wow I’m reading Hacker’s Delight and was just wondering if something like this existed.
275.
by gfodor 4y ago
I listed two reasons: scalability and centralization. You ignored the latter, which is relevant to users today from a pure immediate value proposition standpoint, but people do actually adopt technologies sometimes not based on their curren
276.
by gfodor 4y ago
The author asked why you’d choose this over Mastodon - the answer is that the AT protocol is designed to be a much better solve for decentralized social media that avoids the scalability and centralization problems of mastodon. That’s the e
277.
by gfodor 4y ago
This perspective is too reductionist - we make predictions of success all the time based on first principles reasoning. It’s perfectly sane to try to predict, and make good arguments, if AGI is possible to achieve without new breakthroughs.
278.
by gfodor 4y ago
You’re confusing the goal - the goal here isn’t about the finish line but the point where people all concede that the finish line is actually reachable without any major, presently unthinkable advances.
279.
by gfodor 4y ago
Read the abstract - makes sense to me.
280.
by gfodor 4y ago
I built a thing called webspaces for 3D html worlds based on these principles. https://gfodor.medium.com/rebooting-the-web-in-3d-with-websp...
281.
by gfodor 4y ago
There is no non-AI future. You just get to help choose who gets to whatever the finish line is first.
282.
by gfodor 4y ago
No I’m saying we seemed to live in a world where this was tolerated without incident literally until this week.
283.
by gfodor 4y ago
Understood, but is it reasonable to say that any car sized balloon floating at 20k ft is now worthy of a military response?
284.
by gfodor 4y ago
This looks cool. If I think one step ahead, this reduces enough incidental complexity its the kind of thing that, when combined with AI, could produce the holy grail of an AI writing most of the code for a full CRUD app.
285.
by gfodor 4y ago
That presser kind of blew my mind. I was sure these were obviously ID'ed as balloons by now. And the general seems to have said that we've never shot anything down over the US before, we don't know what these are or where the
286.
by gfodor 4y ago
This was denied as being known by the military presser, and the general involved also made the ridiculous claim that they shot it down entirely to prevent an accident to civilian air traffic.
287.
by gfodor 4y ago
Because it’s being censored.
288.
by gfodor 4y ago
Not sure - this isn’t really a CAD library, it’s more for computer graphics.
289.
by gfodor 4y ago
If you like the idea of a "markdown for 3D models", you should definitely check out Smooth Voxels: https://svox.glitch.me/ I have a fork of it that is faster and also command line tools to make working with these
290.
by gfodor 4y ago
The name of this website is Hacker News. That said it’d be unethical to use this to avoid paying.
291.
by gfodor 4y ago
I don’t assume anything (nice strawman) other than that there are coherent explanations that would explain the presence of non-human technology in our atmosphere that do not require things like FTL travel. If we are going to weigh “extraord
292.
by gfodor 4y ago
Purely irrational bias isn’t really the issue since most people of sound mind can be dissuaded from purely irrational acts. The issue is stereotyping of individuals based upon statistical or anecdotal patterns by people imagined to be a mem
293.
by gfodor 4y ago
The entire question of this thread is the mechanism by which DEI can unwind a given goalpost. (And if/when the last goalpost evaporates, so would DEI.) It sounds to me like your own perspective there is no way for them to do so, given
294.
by gfodor 4y ago
Tribalism is pretty innate but the tribes change. I’d argue if DEI programs had an analog during the reformation era we would be more likely to see Protestant vs Catholic discrimination persist as those programs inflamed it for their own se
295.
by gfodor 4y ago
What you have pointed out is the exact opposite of a falsifiable condition. If we can’t disprove the possibility of a regression back to the past, then this is not meeting the test, and we should expect the results I mentioned. You actually
296.
by gfodor 4y ago
The key thing is there needs to be a falsifiable condition whereby we agree these programs have achieved their primary goal, and that the risk of scope creep, corruption, or other externalities like the ones the OP mentions has exceeded the
297.
by gfodor 4y ago
Correct - it never ceases to amaze me the degree of certainty of many of the so called “skeptics” when it comes to questions for which we have no way to really generate meaningful priors on.
298.
by gfodor 4y ago
Why do you assume the UAPs have aliens in them, and why would you anthropomorphize aliens in any case to be anything like humans? Why do you assume the origin of such objects would be detectable? You're jumping to a lot of conclusions
299.
by gfodor 4y ago
A much simpler explanation than these are objects operated by the US military is that there is a selection bias with public awareness and media reporting on reports regarding the US military. I can think of at least a dozen plausible explan
300.
by gfodor 4y ago
You don't need FTL travel to reach Earth from somewhere else in the Milky Way, if you don't mind waiting a long time to arrive. Which unmanned probes would not. Your confidence in your assertions just demonstrates you haven't

More ›