Support Silicon Dojo at:
RSVP for Classes at:
LinkedIN at: …(read more)
Convert Word 2 PDF: Word to PDF Converter
Convert PowerPoint 2 PDF: PPT to PDF Converter
Convert Excel 2 PDF: Excel to PDF Converter
Convert an Image 2 PDF: Image to PDF Converter
Convert HTML 2 PDF: HTML to PDF Converter
More Tools: PDF Converter
If you were talking about brain operation, your description would basically amount to electric and neurotransmitter signaling. That doesn't explain what intelligence is, nor did your take explain what AI is and what it's supposed to be. One thing is clear, though: LLMs in particular, and ANNs in general, by design mimic our learning process. Given impressive results LLMs are capable of producing, it is incredibly self-assuring to claim LLMs are nothing out of ordinary. The harder will be a wakeup call when it comes.
What if the purpose of those "WormGPT", "FraudGPT" and all these half-backed tools, was only to create the social acceptance for complete control of who has the right to own AI tools ?
So I am into LLMs. Bard is amazing because it is able to cite. It has the power of search and then uses those urls to have footnotes. Most of that is not really the LLM persay it's more of the LLM with a tool.
I enjoyed every second of this video. could not stop myself from blonding-out from start to finish!
I searched through out the whole internet, and I found no one source that is alike Eli the Computer Guy.
I found out about this channel from Louis Rossmann.
Clickbait. Does ChatGPT could teach how to produce, combine, a.k.a. assemble Parts for building 3d printed Glock's, AK's or other Weaponry? they say that Scientists were testing *GPT…
Writing this comment again after seeing another video in feed, forgetting the comment, checking the video and closing the tab 🙂
So what is so different in the statistical token prediction that differs from human speech.
We are fine tuned from early age to speak in our native language(s),
later in life we train new skills by repetition, until our outputs to prompts (questions, dialogue) become acceptable – standardizing.
Brain as an organ functions as a pattern recognizer – LLM training is language pattern training, on scale not achievable to organics on an individual level, but a city with 1 million brains doing the calibrating – now thats some next level computing power, think of trading markets/commodities, how we arrive at a certain price over time – by millions of brains negotiating over prices over sets of time intervals – collective market pricing.
LLM – totally give more interesting conversations than you can have with many people. People's concepts are wrong, so the models trained on faulty concepts are integrated into the token system,
which gives out faulty info.
Dam similar to how we communicate. Ok so the technology is inorganic as supposed to organic, in the end a lot of overlapping between electrical signals/impulses
that run in human bodies and in the transistors of silicon chips – problems of complexity, physics and electrical engineering are overlapping no?
I have heard ChatGBT runs 10×15 to 10×16 computations per second, we humans (brain?) runs 10×18 computations. The simulation with tokens could become indistinguishable when LLM computation power is amplified about 1000x. You have different emergent behavior that you get from a system of 10×15 computation/s , which is different from 10×16, which has new emergent properties nd possibilities when it runs on 10×17 and so forth. – to me this seems like a smart argument.
We have knowledge and expertise from anorganic systems 10×1 to 10×16 – so how much does a 1000 com/s system differ from a sytem of 1 000 000 comp/s and from a 1 000 000 000 comp/s system?
So in my view chatgbt is a simulation of an "average" person, and without knowing it i could relate to it as a person.
Possible faults in my reasoning: humans have also chemical, and light sensitive sensory parts of the cognitive system, organics functioning differs from inorganic in some important ways, military has systems etc that have the technology, expertise etc., LLM is incomparable for some reasons not understood by me, my reasoning is based because of X,Y,Z , consciousness is fundamentally not smth than can be simulated ( although prob you can simulate many people to pretty high degree – our lives are standardised, which means that the best route we can/should take everyday can be put into a flowchart – routine adult life), human speech is different from communicating with word symbols (languagescience diferentiates syllogism, pragmatism, semantics whatever the hell more right so), neuroscience is so complicated that talking bout it is wrong(i imagine the brain approximates something like: 1 neuron x 50 billion + tissue bonding mechanics + DNA filtering/computing),
add here reasons i am wrong ad infinitum ….
I only use internet for porn, I guess I'm safe.
People don’t understand that technology is a tool. All AI needs to be is a data sorter and possibly a teacher w/ a human proctor. Your thoughts on encryption are interesting. Always thought why would you release encryption that you cannot solve for.
Sorry to hear about your wife's health. It just baffles me as to what causes cancer . I have been reading, viewg YouTube videos that if you eat burnt food theburnt part is carcinogenic. Then if you eat processed foods with all sorts of chemicals in it you are at risk.Some types of plastic containers for food packaging has chemicals in it. Smoking, radiation, lining near a radio station or mobile cellular towers increased your risk. We humans are so fragile. However we have a brain to overcome . AI cannot help us overcome
Came here b/c of Louis Rossman.
Instead of joking about how we're all gonna die… Use your technical know how to make a time machine so we can send someone back to save John Connor. Before it's too late. Before… it's… too… late…
Forgive my technical ignorance Eli, but is it possible to tweak one of these LLMs by feeding it a specialized dataset to train it further? Or would you need your own data center with thousands of Nvidia A100s or H100s to train your own LLM from scratch?
The new buzz word that mean everything & nothing at the same time
who is this guy ?
the car and walmart parkinglot data is not threatening the jobs of journalists, WORM is cause if it can convince people to self harm their business with a letter then it can make an article for a news company that will get people to read it,
ie WORM is scary to the journalist because it can replace them. and they can't just learn to code, because the is an AI model that can do that too, better than they will learn how to, thus no job for them, LOLOL
I wonder about the degree to which tech-illiterate members of governments are fooled by this fearmongering bs into keeping AI out of the hands of us plebs
Hello there Eli, the Computer Guy!
I know this may seem petty, but I just wanted to alert you to a spelling error on your webpage under 'What is Silicon Dojo?' | grep 'Hands on Eduction'. Just navigating this website and I love it. Thank you for all that you do.
Sad that you pissed off some dirtbag at youtube, that has to be what happened to your channel.
Mister Snarky strikes again!
Outrageous emails are going to take over the world and turn you into paper clips! #silicondojoforever
Hey Eli, what’s that book on space privatisation called? Thanks
I like chaosGPT
wait what Eli doing jump cuts
Would you be able to do a video on AIOps?
It's interesting how many different ideas and thoughts are complied into one video.
I love this content
Keep it up orange jacket man.
25:30 sad that ppl are scared by llm
20:20 Palantir ?
Linear Regression is officially AI.
I tread where I please computer man, assuming where I please is within a clearly defined and regulated space with prior approval.
Oh wow….an actual informative video…..how dare you ?
wormGPT: "I'm an arsenal."
You had me…., until the Linux bashing started……
"Sad". ~Donald J. Trump
Orange man bad
Ah yes, Computer Man!