How hard is it to build a GPT model truly openly using an approach like SETI / distributed.net using idle times on volunteer computers?
Are there any known initiatives of this kind?
GPT-4
821–830 of 1001 posts
Re: GPT-4
#822After watching the demos I'm convinced that the new context length will have the biggest impact. The ability to dump 32k tokens into a prompt (25,000 words) seems like it will drastically expand the reasoning capability and number of use cases. A doctor can put an entire patient's medical history in the prompt, a lawyer an entire case history, etc. As a professional...why not do this? There's a non-zero chance that i…
> As a professional...why not do this? Unless GPT-4 is running locally on our own computers, there's absolutely no way dumping a patient's entire medical history into this thing could possibly be considered ethical or legal.
> there's absolutely no way dumping a patient's entire medical history into this thing could possibly be considered ethical
Emphasis mine, but isn’t this a rather extreme view to be taking? Ethics deals in the edge cases, after all, so we can easily imagine a scenario where patient consent is obtained and the extra computational analysis provides life-saving insight.
Conversely, the output could mislead the doctor sufficiently to cost the patient their life, so I’m not making any absolute statements either ;)
For the record, and pedantry aside, I do agree with your overall point. Dropping patient history into this thing is incredibly ill-advised. The fact OpenAI retains all your input, including to the API, and provides no low-cost options for privacy is one of the biggest hurdles to major innovation and industry adoption.
Re: GPT-4
#823Let's check out the paper for actual tech details! > Given both the competitive landscape and the safety implications of large-scale models like GPT-4, this report contains no further details about the architecture (including model size), hardware, training compute, dataset construction, training method, or similar. - Open AI
Someone needs to hack into them and release the parameters and code. This knowledge is too precious to be kept secret.
Don't worry. CCP and all kind of malicious state actors already have a copy.
Re: GPT-4
#824Let's check out the paper for actual tech details! > Given both the competitive landscape and the safety implications of large-scale models like GPT-4, this report contains no further details about the architecture (including model size), hardware, training compute, dataset construction, training method, or similar. - Open AI
I've chosen to re-interpret "Open" as in "open the box to release the AI"/"open Pandora's box"/"unleash".
I've chosen to reinterpret it exactly as the kind of Orwellian 1984'ish double-speak that it is.
Re: GPT-4
#825[flagged]
Re: GPT-4
#826If we instead of code feed it AST trees, code that would come out would be epic.
Re: GPT-4
#827It fails on this one, a horse is 15 dollar, a chicken 1 dollar, a egg .25 dollar. I can spend a 100 and i want 100 items total, what is the solution
Using GPT-4 it not only solved this correctly but it generated the linear equations for me:
Let's represent the number of horses as x, the number of chickens as y, and the number of eggs as z. We are given:
The cost of a horse is 15 dollars, a chicken is 1 dollar, and an egg is 25 cents (0.25 dollars). You can spend up to 100 dollars. You want to buy a total of 100 items. We can represent these conditions as a system of linear equations:
15x + y + 0.25z = 100 (Total cost) x + y + z = 100 (Total items) We can solve this system of equations step by step:
Step 1: Solve the second equation for one of the variables, say z. z = 100 - x - y
Step 2: Substitute this expression for z in the first equation: 15x + y + 0.25(100 - x - y) = 100
Step 3: Simplify the equation and solve for y in terms of x. 15x + y + 25 - 0.25x - 0.25y = 100 14.75x + 0.75y = 75
y = (75 - 14.75x) / 0.75
Step 4: Find integer values for x and y that satisfy the equation. We'll loop through possible values of x and check if the corresponding value for y is an integer:
(Deleted some)
Now, we can calculate z using the value of x and y:
z = 100 - x - y = 100 - 0 - 100 = 0
So, you should purchase 0 horses, 100 chickens, and 0 eggs to reach a total of 100 items and spend exactly 100 dollars.
Re: GPT-4
#828Even if a GTP-X will be able to get my description of a system and because it understands the industry lingo and practices create a extremely optimized RUST code, why would we need systems like apps and webapps (like internet bankings) in a world where we all have access to GTP-X?
Like program a KUKA robot to manufacture typewriters in 2023.
How "computer systems" will look like?
Re: GPT-4
#829After watching the demos I'm convinced that the new context length will have the biggest impact. The ability to dump 32k tokens into a prompt (25,000 words) seems like it will drastically expand the reasoning capability and number of use cases. A doctor can put an entire patient's medical history in the prompt, a lawyer an entire case history, etc. As a professional...why not do this? There's a non-zero chance that i…
I must have missed the part when it started doing anything algorithmically. I thought it’s applied statistics, with all the consequences of that. Still a great achievement and super useful tool, but AGI claims really seem exaggerated.
Google: emergent capabilities of large language models
Re: GPT-4
#830I'll be finishing my interventional radiology fellowship this year. I remember in 2016 when Geoffrey Hinton said, "We should stop training radiologists now," the radiology community was aghast and in-denial. My undergrad and masters were in computer science, and I felt, "yes, that's about right." If you were starting a diagnostic radiology residency, including intern year and fellowship, you'd just be finishing now.…
As another radiologist, I'm not sure how you can say this with a straight face? If anything the minimal progress that has been made since Hinton made this claim should be encouraging people to pursue radiology training. As with other areas of medicine that have better AI (interpreting ECGs for example) all this will do is make our lives easier. AI is not an existential threat to radiology (or pathology for that matte…
I totally understand your point #4 - obviously ChatGPT can't do procedures, but I interpreted GP's post as "this is why I did a fellowship in interventional radiology instead of being a (solely) diagnostic radiologist."
But, at the end of the day, diagnostic radiology is about taking an input set of bytes and transforming that to an output set of bytes - that is absolutely what generative AI does excellently. When you said "I'm not sure how you can say this with a straight face?", I couldn't understand if you were talking about now, or what the world will look like in 40 years. Because someone finishing med school now will want to have a career that lasts about 40 years. If anything, I think the present day shortage of radiologists is due to the fact that AI is not there yet, but smart med students can easily see the writing on the wall and see there is a very, very good chance AI will start killing radiology jobs in about 10 years, let alone 40.