lblume

19 ms·

241.
by lblume 2y ago
Well, the code in question is also written by the same LLM, so it could just output something it knows the answers to already. On its own, this result doesn't really seem to prove anything.
242.
by lblume 2y ago
I would argue that the Gamma function is more fundamental than the factorial operation. But you are still correct that if arbitrary functions were allowed, the game would degenerate to triviality.
243.
by lblume 2y ago
So, the formula is really about making any integer with three 2s, but historical precedent calls the game with four 2s more interesting, so the (stronger) result is monkey-patched by replacing a 2 with sqrt(2+2).
244.
by lblume 2y ago
> These percentages are of course tiny and pose no cause for alarm Do people really think like this? Fairly small probabilities of catastrophic events don't generally result in a negligible expected value.
245.
by lblume 2y ago
How exactly does it communicate these changes, if I might ask?
246.
by lblume 2y ago
Additionally, tags naturally form hierarchies in the form of trees (or ADGs), so any possible taxonomy should support that.
247.
by lblume 2y ago
In the first case, I would literally type out the expression faster than thinking about expressing it in natural language.
248.
by lblume 2y ago
The free version, at least once you are logged in, does as well. I don't pay OpenAI, and for the prompt "Calculate the difference between the two biggest primes less than the factorial of 20" it wrote the following code: impo
249.
by lblume 2y ago
Especially for a problem as well-known as this, expect code for it to be seen by the model during training at least a thousand times in different languages, flavors etc. This, on its own, is nothing new.
250.
by lblume 2y ago
Which is a problem with the reputation-based academic system itself ("publish or perish") and not individuals working in it.
251.
by lblume 2y ago
The library of journal templates in the Typst universe is growing day by day. A much bigger issue is some of these journals not yet accepting Typst submissions.
252.
by lblume 2y ago
Typst's file watching seems to work excellent since 0.12, though.
253.
by lblume 2y ago
Sounds like NAND.
254.
by lblume 2y ago
These two are completely different statements, though.
255.
by lblume 2y ago
Think base models. LLMs are extremely good at predicting the future when applied to human languages, as this is literally their only optimization goal. Why couldn't they also be good at predicting the future when applied to other compl
256.
by lblume 2y ago
The environmental impacts of these technologies are well documented and reported. However the emissions are negligible compared to the running emissions of fossil fuels which you don't seem to actively dislike. Assuming that the world&
257.
by lblume 2y ago
Yes, I do. It doesn't distract me a lot, but sometimes my eyes do look at it and it takes a moment to restore the train of thought of what I actually wanted to prompt.
258.
by lblume 2y ago
Then, advanced users should be able to deactivate it in a very simple manner.
259.
by lblume 2y ago
> the people producing and selling solar panels don’t give a fuck about the environment They sure as hell are incentivized to at least make others do. An increased awareness for environmental issues is profitable. I would go further and
260.
by lblume 2y ago
> some people stand to gain a lot from renewable energy program investment Yes, everyone in fact.
261.
by lblume 2y ago
I know that. I read the paper published in Joule, stating that Bitcoin was responsible for 0.2% of global carbon emissions via 0.4% of global energy usage. That, to me, does not sound significant enough to claim it has a measurable impact o
262.
by lblume 2y ago
Do you believe this could realistically be enough pressure to sway politicians driven by corporate financial incentives? Step-by-step reform and gradual carbon reductions just aren't feasible like they have been 20 years ago. Without s
263.
by lblume 2y ago
I would doubt the additional power usage of Blockchain technology and modern AI had a significant impact on global carbon emissions.
264.
by lblume 2y ago
China's exports (and by extension, carbon emissions) are driven by American and European demands for cheap products. Drastically decreasing industrial demand would lead to Asian carbon emissions shrinking by a large margin.
265.
by lblume 2y ago
The question is also for which list lengths the performance matters most. When sorting a few strings (<20), whether the algorithm uses 5 or 7 comparisons would usually not matter too much. So to find the optimal algorithm for a given sit
266.
by lblume 2y ago
Well, of course. Realistically, I would not expect AI systems like this to be very useful for novel cutting-edge scientific results, proving mathematical theorems etc. in the next few years. But this is not the majority of what software dev
267.
by lblume 2y ago
Let's say I can't fully disclose the details because it is an area I am actively working on, but I had an algorithmical problem that was already solved in an ancient paper, but after a few hours of research I could find no open im
268.
by lblume 2y ago
I think that strongly underestimates the impact LLMs, especially reasoning models, have on how code is written today.
269.
by lblume 2y ago
Exactly. Everyone should have the right to have access to this.
270.
by lblume 2y ago
Nim is, in fact, a proper subset of Hackenbush, a game extensively studied by Conway. I found the following video extremely useful and well-visualized: https://www.youtube.com/watch?v=ZYj4NkeGPdM

More ›