I think many people will be quite content not merging with AI and having AI do everything for them. Perhaps only a small subset of people will want to merge with AI. They may not be allowed to though, if majority of people fine not merging then they may ban those who want to merge.
The subset of people who will want to merge will be people who will not be content with having no control over their future. No ability to create new inventions, new discoveries, new ways of working. They will not be content with just having everything provided to them, playing the social games that will remain and think their environment is a prison.
The question is how long before machines can exceed humans in every way and how long will it take before legal and cultural restrictions allow this to happen. It still could be a long time away and jobs will probably increase in the short/medium term as AI unlocks more what humans can do at a rate faster than AI can automate away.
But lets consider what that end state could look like when machines exceed humans in everyway, what good are humans?
One possibility which I don't particularly enjoy is humans will be good for status games, art, creativity, story telling, IRL experiences, everything that involves human to human interaction and connection with other humans. If AI can supply everything in abundance than like rare cards there is only so many humans in the planet at any particular time then human's value increases. I really do believe connections with other humans will have a premium. As a nerd who is borderline anti-social, non artistic, creative, etc this does not appeal at all to me and properly to many others who read HN as well. For the majority of other people with some adaption time I think they will be fine. They were fine adapting to hunters and gathers to agricultural society, they were fine adapting to agricultural society to industrial society, they were find adapting from industrial society to information society and they will will be fine adapting from information society to the social/creative economy. Some short term pain but overall most people will accept this reality quite readily. Those born into it will not even know what we talking about.
This contradiction "we have to make an agreement with China for mutual slowdown" but also "we have to ensure we will always stay ahead of China" would work. Its just fantasy. The only way to enforce this is with military force and the cost would be too unbearable. You could try trade restrictions but the previous tariffs did not work and doubt future ones will to. US citizens (and thus politicians) won't bear that pain. Thus at a minimum this contradiction will mean AI will continue to develop until both US and China are at the same level. Playing with logic gives you possible scenarios. If China catches up within 1 year then negotiations could begin. If China takes many years to catch up, say 6 months behind now, then next year 5 months behind, then year after 4 months behind then expect overall AI development to proceed at full speed. The one way you could do it is to negotiate to transfer direct technology to China that will ensure that they will have the capability to be on par with USA. Politicians won't do this publicly as they want to get elected but they could make a deal that appears to appeal to both sides while actually transferring technology.
I think the only way they could actually pace the frontier is if they restrict the number of GPUs and the power of GPUs available to each person/organization. Licenses would be needed for GPUs above a certain power or equivalent in terms of number of GPUs. Thus, everyone will have to register how many GPUs they have. To enforce this countries would have to use mass surveillance (using AI) on their citizens to ensure compliance as the technology is easier to develop than say nuclear weapons. Next stronger countries capable of having advance AI won't be able to trust weaker countries as they don't know how they will use the GPUs they receive (or even build). Thus like nuclear non-proliferation strong countries will ban weaker countries from having powerful GPUs. If you from a weaker country then too bad for you.
I think AI companies making a point is not the only thing at play. Discovering new maths ultimately leads to new technologies and applications. It may start theoretically but end up being of practical use in the future. Even if humans do not understand it (lose interest, too complex, or just way too many new proofs to go though) AI can use this AI derived math corpus which will help it in other fields.
I would expect this for service providers that provide their "free" services but the mathematicians said they paid OpenAI to help generate this result. This is a clear conflict of interest. Similar to inside trading. Or the same lawyer being hired from two opposing sides which is a big no no. There are rules and laws that already deal with this in other industries and I expect that there will have to be regulation developed for cases where there is a conflict between AI customers and AI provider interest.
Also add corporations. Insurance companies, Banks, Potential Employers, etc will definitively use mass surveillance and AI generated predictions based on that mass surveillance.
Wouldn't the better way be is to make social media open? i.e. anyone can generate their own algorithm or chose from others who created them. I see that the social media companies could implement these requirements in the worse possible way to ensure that they opt right back in to "choose to opt in to having their default feed include personalised content recommended by the algorithm"
From another country but this is the excuse they use here with some paraphasing. "We have decided not to proceed with your complaint ... It looks like your complaint was generated by AI, it is large in volume and confusing. This is the final decision." This was after confirming every single law was valid, referenced, previous case history also properly referenced, added our points, etc. The funny thing is that they only mentioned AI use as a reason to terminate the investigation after many back and forward emails discussing points that they could understand quite well and we pushed back on. Our last response proved beyond a doubt that the opposite party (also another government department) lied to them. For another complaint for a another government agency straight away told us not use AI and use simple words and not to quote the law as that is not how its done in this country. However, we won that one right before telling them we would self represent in court after many back and forward emails with them saying don't use AI each time.
From what I see multiple government departments are already pushing back from people using AI when dealing with the government. I understand the dynamics, if the cost of exercising your rights go down, more people do it and their budgets cannot cope. However, you then abuse the rights of the citizens. Regulations are designed to protect the most vocal people or the wealthy. The most vocal people can start a revolution (or get another party elected) so need to please them while the wealthy are the ones in power. I don't know how this will play out.
My understanding is that you get back a percentage from SynthID regarding how confident it thinks the text is generated by AI or not. The question is what percentage would be considered undeniable that it is AI generated or require more than just your testimony to prove that you actually did the work yourself. https://www.nature.com/articles/s41586-024-08025-4 (Google Deepmind) was using a "Watermark detectability is measured using the true-positive rate (TPR) when the false-positive rate (FPR) is set to 1%" for instance in their research paper. This 1% was set arbitrary just for the research paper but what percentage should it be set to?
Wheres the limit. Would a prompt "Create a system that does x" and no other feedback count? Or do need to do 10 prompts to steer your agent to the results you want. What about just 1 additional prompt. Somewhere there will be a line.
Perhaps but what about individual works done not via a company. Say you are a author and someone falsely detects a AI watermark in your book. Is the testimony you provide enough to override that falsely detected AI watermark? My feeling is that there will be an industry that arises that proves you contributed to the works you want copyrighted.
I thought translations can be copyrighted separately. Their are translators for instance who translate very old texts from other countries/languages and then sell that book/translation using copyright to protect this business models? Not sure but does this mean if you as a human wrote a book in English and then used AI to translate to another language say French then that translated work would not be copyrighted. Not sure how this would work.
As software, creative works, science, etc become more and more contributed via AI does that mean all future works will not be copyrighted or patented? Under our current copyright culture and laws obviously not but that does open the question to how much does a human have to contribute and what evidence is required to show that a human contributed enough so that it can be copyrighted and patented. Some time in the future (may be many years) science may become to complicated for humans to understand fully with AI systems researching themselves. Current patent laws in many places including US says inventions created entirely by artificial intelligence cannot be patented. This future may be coming. What will the new copyright and patent laws look like in the future? Do we need copyrights and patents anymore?
Agree with programming is meditative but disagree that agent-based development means there is no thinking. Agent-based development means more thinking like a lower level manager. You spend more time making architecture decisions, making decisions on the user interface, trying to manage your time and your agents time to increase the amount of work produced in the same time, asking the agent about the code and making new decisions. For many people who went in programming thats not their strong suite and hence is really exhausting. Instead of living in the code you are managing the code and using a part of the brain you are not good at.
The other side is that I have witnessed doctors making mistakes personally on medical files. They are overworked and only see you for 15 minutes and in a rush to kick you out and so of course they make errors. It would be interesting to compare whether transcription software or humans make the most mistakes. Guessing the answer will be very nuanced such as transcription software failing for accents not trained on, how over worked the clinic is etc.
Its not just sleeping more but having more people wake up in their optimal times. People's natural wake up times are in a distribution so when you remove the commute time you including more people in the optimum wake up times.
interestingly moving between jobs is more likely to increase workers salaries. Would be interested if one of the reasons stopping wfh is to prevent salaries from increasing. There is a physical barrier to finding new employers that are further away from where you live and less jobs available overall in the area you live compared to the entire country/world. People who wfh can work almost anywhere, do not have a moving barrier and can job hop their way to higher salaries
Some camera pointing at you while you work. Guessing the camera needs to be watermarked itself, which can be done by adding some watermark on the chip level (each camera will have different watermark), manufacturer can confirm the watermark.
Now for the human generated watermark. Timez to addd the speelling mistakes, decreaze the usegage of big words and proper gramicaly usuage. Wish I was joking.
Worse, what will academic institutions decide is the threshold for detecting AI generated work. If you have a false positive how do you prove it was a false positive or we all just trust the watermark detector over the student saying "I swear I did it all by my self"
As usual a balance is needed. There is also a concern that society (friends, colleagues) that always shuts down your ideas just because they are different thus reducing creativity and progress. Sometimes you need a way to develop your ideas without the outright rejection and sometimes you need the filter of society to see if the ideas really hold up. Usually the time where ideas are most fragile to being destroyed is when they are just developed. Perfect time to brain storm and increase the strength of those ideas with AI before taking it to the filter of society.
Same. I don't remember most things including my childhood. I have to relearn everything all the time just in a few months time. Maths, code, etc if I don't use it in 1 or more month, I have to relearn it though its always easier to relearn it. Comparing myself to others its extreme. What I find is the process of constantly relearning instead of relying on memories (which are in a way similar to assumptions) can lead to different ideas (often unique) than other people because everything even basic concepts is a open question. This however is tortuous in job interviews when people ask about problems solved in previous work or about a concept so there are trade offs.
I suspect any new thing created/discovered/etc is just combining knowledge from existing things with memories being an example. Outremembering while certainly not the only method helps massively with combining knowledge from existing things. I don't see new knowledge coming out of nothing.
Realized your right about OpenAI but they plan to. In the link they state "our goal is to expand provenance signals to all modalities including text" so its comming up regardless. I pretty sure Google does it though.
Not sure why Anthropic is getting all the attention here. Google been doing this since at least late 2024[1] and openai I believe is doing this around 9 days ago[2].
If a computer vision model can identify ads and draw a rectangle over them, it can identify the DOM elements rendered inside that rectangle. Thus, the HTML, JS, and network calls of those ads can be detected. This data can then be used to train an ad detector model based on the DOM and network examples. The computer vision model essentially trains the DOM/network blocker, acting as a fallback mechanism when the blocker fails and continuously retraining the ad detector model.