namaria

9 ms·

61.
by namaria 1y ago
> The Dawkinsian selfish gene framing is unfalsifiable. Disagree. The whole thesis rests on genetic determinism, which was hypothesized early on - around the time the book was written - but discredited later. > Even Darwin is practica
62.
by namaria 1y ago
The anxiety is there whether you examine it or not. Processing things that give us anxiety is the way to get rid of it.
63.
by namaria 1y ago
Meditation is such a superpower. It costs nothing - not even time when you realize through meditation that there is no such thing as time, and emptying your mind is a great way to be in the now - it's "easy" in the sense that
64.
by namaria 1y ago
I find myself lacking the necessary mental down time to process things that happen in my life. Long walks help. No music or podcasts, just me, my mind and the city.
65.
by namaria 1y ago
Hope was the only evil left inside Pandora's box upon opening. I've always wondered about that. Giving hope is the ultimate evil, "too much" even for a divine punishment?
66.
by namaria 1y ago
> It took a solid hundred years to legitimate photography as an artistic medium, right? Not really. "In 1853 the Photographic Society, parent of the present Royal Photographic Society, was formed in London, and in the following year
67.
by namaria 1y ago
Necessary != sufficient.
68.
by namaria 1y ago
In the very narrow fields where I have a deep understanding, LLM output is mostly garbage. It sounds plausible but doesn't stand up to scrutiny. The basics that it can regurgitate from wikipedia sound mostly fine but they are already s
69.
by namaria 1y ago
Just using LLMs on the (few) things I have specialist knowledge of it's clear they are extremely limited. I get absurdly basic mistakes and I am very wary of even reading LLM output about topics I don't command. It's easy to
70.
by namaria 1y ago
Limited blast radii are a great advantage of deterministic tools.
71.
by namaria 1y ago
To be fair, I find ChatGPT useful for Elixir, which is pretty niche. The great error messages (if a bit verbose) and the atomic nature of functions in a functional language goes with the grain of LLMs I think. Still, at most I get it to hel
72.
by namaria 1y ago
Does it sometimes instead of driving a nail hit random things in the house?
73.
by namaria 1y ago
I will walk around the gratuitous rudeness and state the obvious: No, the pretend above 95% accuracy is not as good as the up to 50% rates of hallucinations reported by OpenAI for example. The difference in experiences is easily explainable
74.
by namaria 1y ago
I just want to have a dignified journey. Help those around me, be a little bit better everyday, and leave peacefully when it's time. Hoarding anything seems like a way to get attached and thus unable to face my curtains time gracefully
75.
by namaria 1y ago
> Now, I could totally see a scenario where if I spent time tweaking prompts, writing rule files, and experimenting with different models, I could improve that output significantly. I think this is it. Some people are willing to invest t
76.
by namaria 1y ago
That we're getting either sycophantic or intransigent hallucinations point to two fundamental limitations: there's no getting rid of hallucinations, and there's a trade-off in observed agreement "behavior". Also, th
77.
by namaria 1y ago
People keep throwing these 95%+ accuracy rates for LLMs in these discussions, but that is nonsense. It's closer to 70%. It's quite terrible. I use LLMs but I never trust them beyond just doing some initial search if I am stumped,
78.
by namaria 1y ago
> the program doesn't compile How does this even make sense when the "agent" is generating Python. There are several ways it can generate code that runs and even does the thing and still has severe issues.
79.
by namaria 1y ago
Y'all starting to sound awfully like religious types. Is that satire?
80.
by namaria 1y ago
That's basically the whole premise of Bostrom's "Superintelligence"
81.
by namaria 1y ago
LSD is a dirtbike of the mind. Some people can do amazing cross country trails, some people can fall off and break their skulls instantly.
82.
by namaria 1y ago
Sure and FTL, if invented, will lead to the colonization of the nearest systems.
83.
by namaria 1y ago
No, that's not the same thing at all. LLMs are termite mounds, not the colony.
84.
by namaria 1y ago
No, we know planetary ecosystems can use energy gradients to sustain intelligent lifeforms. Intelligence is not a feature of the human brain, it's a feature of Earth. Without the ecosystem there are no cells, no organisms, no specializ
85.
by namaria 1y ago
Training LLMs is literally finding sets of numbers that make them better at mimicking human language.
86.
by namaria 1y ago
And if my plane keeps the take-off acceleration up for 7 months we'd be at 95% the speed of light by then.
87.
by namaria 1y ago
Yes. Check the codebases. It's all prompting scaffolding. All of it. Chain of though, agents, tool using. It's just parsing the user inputs, adding text around it. "You are an expert X". That's the whole edifice.
88.
by namaria 1y ago
It's all 'clever prompting' all the way down. 'Chain of thought', excuse me, 'reasoning', 'agents'. It's all just prompt scaffolding and language models. There is no way this ends the way th
89.
by namaria 1y ago
So your argument is: Athens and Sparta had things in common but were different. Like Czechia and Russia. Czechia and Russia are quite different. So were Athens and Sparta? That's called circular reasoning.
90.
by namaria 1y ago
> ... which is my point as well. You can no longer tell if your interlocutor is even human I don't really think modern chatbots pass the Turing test. It's not that hard to figure it out. > (NLP wasn't much of a money pi

More ›