MaidSafe using Ai?

I heard MaiSafe is getting closer to launch?

Is that a consensus of any type 'round these parts?

Ai is taking the world by storm, is CoPilot X helping SAFE development at all? Or did it help the UDP hole punching team achieve their goals perhaps lately? I know that was a major step forward for everyone.

What are your guys thoughts?
I like to come and check in time to time.

Wonder if it will be us lovely humans of earth and Ayr who make the first autonomous decentralized network, or if strong AI will be developed first, and create SAFE before us humans can :joy:

bets?

(Great to see SafeCoin being RFCā€™d up and discussed more and more lately btw, thatā€™s a huge driver of real, self-sustaining, living economy, true SAFE launch)

6 Likes

Some have been using copilot and gpt3/4 for a while.
Several of us are starting to use it more and more now, for various sorts of problems.

13 Likes

Does Europe have GPT premium (subscription) now? I know it was only available in the US to start.

Yeah, I know at least myself and David are subscribers.

5 Likes

Gpt plus is actually ā€œout of stockā€ so many people that gpt plus is still lagging or out of reach for most, you need to use it when most people are asleep

1 Like

Game changer for sure, worth every penny. Nothing more frustrating than the limits imposed on the free version. Iā€™d cancel other subscriptions to keep this if needed.

3 Likes

Yeah, to be honest, I was glad to see them introducing a subscription service early as an alternative way of funding. I donā€™t want to see it turned into an ad-ridden shitstorm.

Even as a subscriber, the performance of GPT-4 is quite poor. Most of the time if you have an answer that requires a long response, it gets truncated. For short responses, it works ok, although a little slow, but thatā€™s not really a big deal.

GPT-3/3.5 is lightning fast a subscriber though, and still very useful for probably most things.

5 Likes

I wish greek doctors in hospitals use ai soon enoughā€¦ I am with a buddy here in emergency room and we are here 7 hours here and still waiting cause there are patients coming with ambulance and need immediate helpā€¦ One of them dead in the doctors office taken out covered

Also some drama , a body builder came and was shouting, talked to the head of the emergency and he was treated immediatly, apparently he didnt sleep for 2 nights straight but again if he didnt pump iron he would be advised to wait like all of us

4 Likes

In a very general sense, at the moment, I have mixed thoughts about it.

There is no doubt that the technology is incredible, and Iā€™m generally for AI replacing humans for all the important aspects of our societies; healthcare, as mentioned, being just one example.

Though I also feel it makes for a disconcertingly uncertain future. Iā€™ve been interested in discussions about ā€˜futurismā€™ for around a decade or so. People in that space have been talking about AI for a while, but for me, ChatGPT changed it from an abstraction into a stark reality. The moment it was released, effectively we entered into a transition period for industries like IT, in their current form, to finish. Iā€™m not talking tomorrow, but over the next five-to-ten years. Whether these industries will turn into something better, or lead to mass displacement and vastly more inequality, is the issue.

Itā€™s an unstoppable arms race now and we just need to see what happens. It doesnā€™t feel good to have no control over that, but not much you can do I guess.

10 Likes

There is an open source gpt alike that i read its up there. Cannot recall the name or website but i will search for it cause it changes the discussion around ai nad how closed garden it is

1 Like

I read a lot of stories on reddit on gpt sub that has helped people find out a possible diagnosis, and most of the times it is proved accurate. So i guess doctors or any profession can put all the data on chatgpt and then just verify the results. I have read of one person giving large data to chtagpt and replacing a 10k$ consultant

1 Like
10 Likes

Yeah, I mean, personally I would rather get a medical diagnosis from a machine that has read literally every piece of medical literature ever written, rather than a very fallible human. (Although itā€™s a bit crushing to get a bad diagnosis from the cold perspective of a machine!)

But it does lead to the question of where it leaves the humans who currently make a living doing thatā€¦I think there are potentially going to be a lot of uncomfortable changes for people. I really want to believe there are going to be these ā€œbetter jobsā€ that people like Sam Altman speak of, but itā€™s not clear.

7 Likes

I find that ā€œPlease continueā€ works well if you get an answer that seems truncated. EDIT on ChatGPT - I havent tried the upgrade yet

2 Likes

I donā€™t disagree with what you say, but Iā€™m also aware that thereā€™s a problem with that medical model. Well more than one problem.

Weā€™ve moved from a system where when I was a child a UK GP knew all their patients and I was visited at home for minor illness. Now the model is one of symptoms, diagnosis and treatment which is fundamentally limited.

It already largely ignores causes, local and interpersonal knowledge, and crucially the importance of a personā€™s relationships, intimacy and community in their life

One of the biggest determinant factors in health (immunity, survival, recovery etc) is social. Automating the Western medical process undermines that rather than improves it.

Thereā€™s a pull towards a certain use of medical AI (efficiency, accuracy etc) which misses a lot IMO, and is ultimately unhealthy.

So while I agree, and would make use of medical AI I believe I can do a lot myself that is also influential by working towards a more connected life, with nature, animals, friends, community etc.

6 Likes

As well as all those shortcomings, there is also the strictly for-profit Yankeeisation of medical care . .

3 Likes

Although using chatgpt for medical diagnosis bypass the privacy protections we have in the medical system in many if not most countries. And being a subscriber it is tied to you. Even subscribers are still being treated as the product by chatgpt where you info is available to be used elsewhere.

That was the same as here. But we still go to the GP of our choice and medicare pays the same for any doctor and the patient pays a gap if the doctor charges a higher rate.

What Iā€™d like to see is a combination of going to GP of choice and he knows your history plus using an AI to assist. With the additional knowledge of history the GP can improve the input to the AI and also many times not even need it since some things are obvious to the trained eye.

3 Likes

They would also act as an anonymizing layer. Naturally a doctor (or hospital, or any level) will have many many patients, and would input no data that would connect to an individual.

3 Likes

I would expect that within 5 years there will be a specific one for medical and able to run on a local machine to the surgery. And also the laws will have caught up such that the protections on medical data will apply to any AI making it illegal to access that data by anyone other than the GP consulting or hospital treating the patient.

4 Likes

I think in the future we are going to have two types of AIs - highly specialized with all available knowledge and high reliability in narrow field (medical diagnosis, accounting consultant,ā€¦), and the second type would personal assistant which would be under personal control, probably run locally on phone or home server. Personal assistant would be the one you communicate wit, it will communicate on your behalf with the specialized AIs, choose the one best suited to do the job, pay for their services etc.