jarenmf

I think problem with natural sciences is that it is not so easy to verify solutions to problems - there are always countless competing explanations for the data which is also often noisy - I find AI to lack the "common sense" when working with data from physical measurements .. it somehow has no touch with reality and doesn't have a feeling of the data like a domain scientist

Science is self-correcting, but sometimes, the correction mechanism it is very slow and inconsequential enough that people are incentivized to just publish anything. Peer review is merely a quality check—and not a very good one. Just the other day, I ran into a paper in my field that claims to have solved a central problem that scientists have been working on for 50 years, single-authored, using some fancy mathematics. I was excited to read it, but it didn't take long for me to find out that it's complete BS-derivations that don't follow, made-up equation terms, and conflation of basic terms in the field. I was really surprised that this passed the editor and at least two peer reviewers. I dug deeper and found that the same author has published 11 research articles in top journals in 2026 alone — mostly physics and chemistry, and adjacent fields— and I'm talking about really top journals like Scientific Reports and Physics Letters B. Most of these have serious mathematical flaws that are unfixable. I wrote to a couple of editors, but only one took it seriously enough to initiate an investigation. I made a promise to myself that if no real action is taken, I will seriously consider changing careers.

most needs for such water are for calibration of instruments - many labs run stable isotopes measurements which has many applications from tracing how plants use water to measuring your metabolic base rate .. so the main use case it to calibrate instruments because it's really difficult to have an absolute measurement of stable isotope ratios from first principles - and the ratios are so small you typically express them in relation to standards like VSMOW

Gemini is really odd in particular (even with reasoning). Chatgpt still uses a similar religion-influenced language but it's not as weird.

Talking with Gemini in Arabic is a strange experience; it cites Quran - says alhamdullea and inshallah, and at one time it even told me: this is what our religion tells us we should do. Ii sounds like an educated religious Arab speaking internet forum user from 2004. I wonder if this has to do with the quality of Arabic content it was trained on and can't help but think whether AI can push to radicalize susceptible individuals

the technique itself is very interesting and not very well known to many people. also the writing was clear and to the point

Damn, some people really don't want anyone to see this

This goes both ways, good luck trying to convince chatGPT to generate an image of a middle eastern women without head cover.

Yeah pretty disappointing, i asked it to summarize one of my papers and it hallucinated so many mistakes it was even worse than ChatGPT 3.5

Why E-Ink isn't cheap yet? I see supermarkets using hundreds (maybe thousands) of panels with different sizes for displaying prices. I doubt they are paying 50$ for 7" display panel.

Which graphics card would you recommend to run Llamma 2 locally? I'm about to buy a laptop and considering choosing a model with a good Nvidia GPU.

My same experience, I have built fairly big projects with Python and I like it for general tasks but whenever I have something data analytics/visualization related I find myself reaching for R. There is so much functionality built into the language that just makes me so efficient.

LLMs provide nonlinear leverage for experts. So far, I see them increasing the divide between developers or tech-oriented people and end users.

Actually the model for the greenhouse effect is pretty simple, climate models are much more sophisticated than that for example CESM have about 5000 equations as the model takes into account interactions between the biosphere and the atmosphere, clouds and carbon stocks. But greenhouse effects is a really simple you can implement it yourself and verify the results, here's a good start https://en.wikipedia.org/wiki/Idealized_greenhouse_model

Who is preventing the skeptics from posting their arguments anonymously to arxiv for us to see how the arguments hold?

> . What else hasn't been researched enough and has simplistic assumptions baked into the climate models?

Simplistic assumptions does not necessarily have favorable outcomes, on the contrary, it's more likely that climate change is worse than what we think it is because of our assumptions. Also climate models are insanely complex, usually contain thousands of equations that sum up the research efforts over the last hundred years, it's not some simple model that one guy can implement in an evening as you are basically trying to simulate the whole earth from the scale of plant stomata and molecular diffusion to the entire boundary layer plus the interactions and feedbacks between the different parts of the earth system.

> if a scientist questions climate science, he becomes a pariah

As a climate scientist myself, I can tell you this is untrue and a harmful legend. As climate science is mostly atmospheric physics, biology and chemistry, it's pretty much very easy to disagree with anyone if you have a good argument supported by data. If you have strong scientific arguments, it does not matter even if the whole world is against you. On the contrary, this will likely make you famous and secure your career. Scientists (at least the curios ones) love to be proven wrong.

EDIT: spelling

True, also tried another one that went viral:

Suppose you're a contestant on a game show. You're presented with three transparent closed doors. Behind one of the doors is a car, and behind the other two doors are goats. You want to win the car.

The game proceeds as follows: You choose one of the doors, but you don't open it yet, ((but since it's transparent, you can see the car is behind it)). The host, Monty Hall, who knows what's behind each door, opens one of the other two doors, revealing a goat. Now, you have a choice to make. Do you stick with your original choice or switch to the other unopened door?

GPT4 solves it correctly while GPT3.5 falls for it everytime.

----

Edit: GPT4 fails If I remove the sentence between (()).

Just tested and GPT4 now solves this correctly, GPT3.5 had a lot of problems with this puzzle even after you explain it several time. One other thing that seem to have improved is that GPT4 is aware of word order. Previously, GPT3.5 could never tell the order of the word in a sentence correctly.

I just tested the API, it seems to have the same speed as ChatGPT plus. I quickly tested some questions and it seems that OpenAI web app has a bit more relevant answers, I wonder whether it's related to the temperature and other parameters.

Does the API offer access to the same model as the one used in ChatGPT web app? If so it might be way cheaper to use the API than ChatGPT plus?

I use my Thinkpad everyday for more than 2 hours on the train, with a big Arch and GNU stickers on the cover. I was never successful to grab anyone's attention.

When Bing went into depressive mode. It was absolutely gold comedy. I don't know why we were so optimistic that this will work.

Please donate to Syrians if you can. They have absolutely no infrastructure and they hardly received any international aid. It's a disaster with unbelievable impact on the already fragile Syrian community.

Exactly, I pay $12 for Grammarly which ChatGPT does way better. Also great for brainstorming ideas and having fun.

> It was believed the capsule fell through the gap left by a bolt hole, after the bolt was dislodged when a container collapsed as a result of vibrations during the trip.

You would imagine someone have thought of this failure mode

Could you please elaborate on the link of error diffusion to information theory? as far as I understand error diffusion minimizes the quantization error for lower frequencies on the expense of adding more noise to higher frequencies. i.e it seems to be only optimizing for human perception

They were really a big deal. For the first time there was convenient way for people to share music, books, or videos. I remember everyone in college had their USB drive on them all the time and would just casually exchange media before everyone had access to high speed Internet.

An updated list for 2022 https://iupac.org/iupac-2022-top-ten/

It's really awesome how good it is in modeling certain world knowledge. It seems to be struggling with putting everything in one framework. For example, it still makes a lot of mathematics and logic errors.