gibsonf1

8 ms·

61.
by gibsonf1 2y ago
Yes, but it will hallucinate like all other LLM tech making it fully unreliable for anything mission critical. You literally need to know the answer to validate the output, because if you don't, you won't know if output is true o
62.
by gibsonf1 2y ago
I hope all the LLM features can be disabled.
63.
by gibsonf1 2y ago
Yes, but what about all the bugs created and time to debug?
64.
by gibsonf1 2y ago
Actually, its worse than that. They froze Starlink as Brazilians can bypass state internet censorship via that. Its all about censorship.
65.
by gibsonf1 2y ago
This looks fantastic!
66.
by gibsonf1 2y ago
Hmm, I would not say that a change in state = time, but that a state changes at some point in time where the state is current for the period of time (start to end) in which some measurable attribute of an object does not change. The standa
67.
by gibsonf1 2y ago
Yes, basically as long as you keep measuring the object with a function (like length or weight etc) and the result of the measurement does not change, the state has not changed. When the result changes, its a new state.
68.
by gibsonf1 2y ago
Actually, a state is when a measurable attribute of an object remains unchanged for a period of time. So saying state = time completely misses what a state is.
69.
by gibsonf1 2y ago
Ability to write at a very high level (and macros), incredibly fast when compiled, ability to use the repl on a live server for diagnosis and patching functions, language design choices work extremely well for parallel recursive hierarchica
70.
by gibsonf1 2y ago
Conceptual AI: https://graphmetrix.com/ (Common Lisp)
71.
by gibsonf1 2y ago
This is great news. The updates to SBCL broke some libraries (that is, sbcl got more thorough in finding bugs that some libraries had) and I've had to load patches in for them because of the quicklisp limitation in a patch file. This
72.
by gibsonf1 2y ago
Given that GenAI is a statistical approach from which intelligence does not emerge as ample experience proves, does this new company plan to take a more human approach to simulating intelligence instead?
73.
Brands Are Beginning to Turn Against AI(rollingstone.com)
5 points by gibsonf1 2y ago|3 comments
74.
by gibsonf1 2y ago
Isn't 50% kind of a failing grade?
75.
by gibsonf1 2y ago
Exactly my thought, at a minimum, it should match the speed of the original.
76.
by gibsonf1 2y ago
https://graphmetrix.com/trinpod-server
77.
by gibsonf1 2y ago
The "degraded" performance in our case seems to be all api requests.
78.
Wasabi non-operational us-east-1 and us-east-2(status.wasabi.com)
2 points by gibsonf1 2y ago|1 comments
79.
by gibsonf1 2y ago
I haven't heard of any Orca success with either skeg-hung or long keel boats. A big "new" development in sailboat design is dropping the skeg on rudders for better sailing performance and lower cost even on blue-water sail b
80.
by gibsonf1 2y ago
Being a sailor, I've been carefully paying attention to the Orca reports, they seem to like the modern sail boat fin rudders which have no skeg (vertical support attached to hull with hinged rudder behind) and is pretty easy to knock o
81.
by gibsonf1 2y ago
Except you’d be missing the part that a neuron is not just a node with a number but a computational system itself.
82.
by gibsonf1 2y ago
That looks insanely expensive.
83.
by gibsonf1 3y ago
A better way to do this would be to more closely model what is really happening, that is the transaction events which change state. The states being changed would be the account balance amounts input to and output from the transaction even
84.
by gibsonf1 3y ago
I would say conceptual awareness is a far bigger bottleneck than visual perception data.
85.
by gibsonf1 3y ago
I find it interesting that they promote the idea that they are pursuing AGI when they don't even have any I yet in their products.
86.
Google CEO Calls Biased AI Chatbot Responses Unacceptable(wsj.com)
22 points by gibsonf1 3y ago|15 comments
87.
by gibsonf1 3y ago
I'm not arguing that they are not fantastic statistics machines, but there is nothing like semantics or "two concepts" for LLM, its just word strings that statistically occur together, etc. This is precisely why they are so
88.
by gibsonf1 3y ago
My basis would be the conceptual AI on top of space-time my company has built: https://graphmetrix.com/ I started the epistemological research in the late 90's for understanding how human concepts work and started buil
89.
by gibsonf1 3y ago
Sure we do. When you read a sentence, and are able to model that in your mind and understand the space-time being referred to, you understand it. LLM is like a calculator in the sense that given an input it gives an output. There is no m
90.
by gibsonf1 3y ago
I'm not sure what you mean by the system learning "perfectly well" as it literally knows nothing about what those words mean. This is why LLMs can give statistical word results that can be true of false or in between and tho

More ›