Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror

Comment Re:Slam dunk case (Score 0) 108

we are on a technical site, supposedly. The work required to ensure that every single seat is marked 'window but not really' vs 'window' and then all of the plane schematics are adjusted according to every single airplane is insane and may not even be technically possible. You are buying tickets through a system that doesn't know the exact layout of that particular airplane, it knows the model. The layouts change because airlines replace seats and may change the number of rows and how the rows are aligned exactly against windows and doors, etc.

I am saying this is ridiculous to sue for this, because it's clearly way too complex of an issue, just like most other real life issues, they are not 'black' and 'white', 'democrats' and 'republicans', etc.

Comment Re:Slam dunk case (Score 0) 108

This is nonsense, what if it is not a window but a door instead, should you be suing based on a technical name? Window seat is seat adjacent to a wall, aisle seat is a seat adjacent to a corridor. Then there are middle seats. I got 'window seat' many times, where the window is not directly in my row but it is somewhat between rows, so what? It's a name.

Comment Alaska (Score -1) 19

If ruzzian terrorist putin arrives to Alaska and instead of immediately arresting and throwing him to jail and then offering ruzzia to start actual negotiations Trump meets with him to discuss anything at all, I believe this will mark the end of the USA as a nation altogether. You can then forget about your very concept of laws, don't bother, there is nothing, just emptiness and pointlesness.

Comment Will the AI crash lead to another AI winter? (Score 1) 238

"... the apparent reasoning prowess of Chain-of-Thought (CoT) is largely a brittle mirage. The findings across task, length, and format generalization experiments converge on a conclusion: CoT is not a mechanism for genuine logical inference but rather a sophisticated form of structured pattern matching."

The latter should really not surprise anyone with a passing understanding of the LLM transformer model. They were never designed for reasoning tasks but for machine translation. But an entire industry has now sprung up trying to shoehorn them into arbitrary business cases, no matter what level of real reasoning, expertise, common sense and judgement is required. I am quite confident in predicting that Sam Altman's quote that "GPT-5 is the first time that it really feels like talking to an expert in any topic, like a PhD-level expert," will live in infamy.

Comment Re:chess (Score 1) 238

I did say this though:

couldn't redraw the board from the list of moves, I pasted all of the moves, it couldn't draw an accurate board, basically fell apart on me after a little while.

I did try to 'remind' the LLM what it was supposed to know anyway, it couldn't work even from the newly provided list.

Comment chess (Score 2, Interesting) 238

I understand that humans can no longer beat chess engines, I am not a good player, I dabble. I also understand that LLMs have no real memory or game state, etc. Asked ChatGPT5 to play a game yesterday, I set up a physical board, it drew ascii board and used annotations. For a little while it was OK, maybe the first 15 moves or so. It wasn't beating me, it was balanced at first, however it started behaving as a drunk would. Forgot how pieces move, forgot where some pieces were, forgot that it was white and would tell me that it is white's move instead, couldn't redraw the board from the list of moves, I pasted all of the moves, it couldn't draw an accurate board, basically fell apart on me after a little while. When things get too complex and require either actual thinking or specialized algorithms and databases it failed. It is not intelligence, just a poor simulation.

Comment Re: Sold his stock (Score 0) 98

imagine, I don't surround myself with employees who would throw a temper tantrum when presented with any challenge. They can try and fail, they can present an opposite scenario or whatever, as long as they don't go ape shit crazy I work with them. I have fired a few people over the last decade who failed at this.

Comment Re: Sold his stock (Score 0) 98

See, I wouldn't hire you, you just went ape shit over absolutely nothing. First of all I never ask really irrelevant questions in my interviews, however I do pose scenarios that are obviously linked just to the job description itself, I want to see if the candidate can handle himself and how he thinks, more importantly I want to see him or her handle the interview.

Slashdot Top Deals

HOST SYSTEM NOT RESPONDING, PROBABLY DOWN. DO YOU WANT TO WAIT? (Y/N)

Working...