threatripper

9 ms·

91.
by threatripper 2y ago
Can't we tokenize numbers always as single digits and give the LLM an <thinking> scratchpad invisible to the user?
92.
by threatripper 2y ago
That was during the great plugin times. Haven't seen plugins for a while. Do they still exist?
93.
by threatripper 2y ago
Fascinating how some road paths have endured the centuries.
94.
by threatripper 2y ago
Here they talk about the solutions developed in China: https://www.youtube.com/watch?v=_1f-o0nqpEI Basically they made training much more efficient and developed a better and cheaper product. They passed the stage of "
95.
by threatripper 2y ago
It always matters to some degree but you can spend much much more to get an edge for national security. 10x maybe 100x compared to the free market but not infinite.
96.
by threatripper 2y ago
Even if price and energy is 10x it will not slow down AI development much once the Chinese government starts pouring in big money.
97.
by threatripper 2y ago
Those are two different sentences. The second sentence doesn't refer to experts explicitly.
98.
by threatripper 2y ago
Requiring only simple prompts surely sells better. I would not assume the documentation provided by OpenAI is totally unbiased and independent of business goals.
99.
by threatripper 2y ago
The basic question for the next years is if hydrogen from fossil fuels makes sense. There's no other way to make enough hydrogen in the short term. In the long term I think that batteries will be better suited to use renewable electric
100.
by threatripper 2y ago
For him and his boss and the boss of his boss it may well be a goal to use more AI in business processes. It may be decided in the strategy to spend X% on AI in the next 3 years. So you will do exactly that and not question if it makes sens
101.
by threatripper 2y ago
Or 10x the skill and speed of a human in some specific class of recurrent tasks. We don't need full super-human AGI for AI to become economically viable.
102.
by threatripper 2y ago
Paywalled article for me.
103.
by threatripper 2y ago
You take a picture of a chess board and send it to ChatGPT and it replies with the current evaluation and the best move/strategy for black and white.
104.
by threatripper 2y ago
I'm pretty sure that LLMs already overpower the average arguing abilities of users on many platforms. Not even low resolution video calls are safe anymore to verify somebody is a real person.
105.
by threatripper 2y ago
Joules per character
106.
by threatripper 2y ago
Humans consume a lot of power and resources.
107.
by threatripper 2y ago
Has this been investigated in detail by physicists? Does it hold up in theory?
108.
by threatripper 2y ago
Don't you think that using a time format strongly tied to a religious calendar and a operating system with a distinct philosophy is a bit problematic? GPS time is a bit heliocentric and imperialistic but certainly more free than UNIX t
109.
by threatripper 2y ago
You should actually ask for 3 examples so you can select one.
110.
by threatripper 2y ago
https://artificialanalysis.ai/leaderboards/models "LLM Leaderboard - Comparison of GPT-4o, Llama 3, Mistral, Gemini and over 30 models Comparison and ranking the performance of over 30 AI models (LLMs) across key
111.
by threatripper 2y ago
Their technology deserves a lot of praise but their anti-competitive business practices should be shamed. If there was a way to separate the two i would love to see intel as a business go bankrupt but intel as a foundry live on.
112.
by threatripper 2y ago
> Moreover, the military wants parts for decades, so you're stuck building obsolete parts on an obsolete process. Intel recently built expertise in this area, so I think they are a good match.
113.
by threatripper 2y ago
That's pretty much it. It's like the nose in your visual area. It's always there but normally you don't see it. Focus on everything else. Let it be there but focus on something.
114.
by threatripper 2y ago
If your inner motivation is satisfying intellectual interest or improving skills, you actually finished the real project behind the facade without a need for finishing the official project. So, the mistake boils down to making up a fake off
115.
by threatripper 2y ago
Is it this one? https://github.com/sungminlee114/PEAQ_matlab
116.
by threatripper 2y ago
This one is an interesting take. In hindsight it's pretty obvious but if you are in that situation at that time it seems like you would be selling your crown jewels for pennies.
117.
by threatripper 2y ago
Have you tried looking at your problem from different standpoints and with different ways of thinking? Have you tried to understand the ways in which your listeners are thinking about things?
118.
by threatripper 2y ago
Now i wonder if there's a connection between j and ; as termination character.
119.
by threatripper 2y ago
I really wonder where the limit is for AI. Reality has an incredible amount of detail that you can't just simulate or emulate entirely. However, our perception is limited, and we can't process all those details. AI only has to be
120.
by threatripper 2y ago
I wonder if we look back at Meta like this which pours ungodly amounts of Money into VR right now.

More ›