Are you one who doesn't believe that AI is changing things?
There are many people who are ignoring the facts. We still have many facets of the population that believes AI will not surpass humans in many fields. To me, this is a viewpoint that is rapidly getting destroyed.
Sakana AI is a Japanese company that caught a lot of attention last year when it claimed to be creating the first AI scientist. It was something that naturally met with skepticism.
Like most things in the AI world, if you do not believe something, wait 6-12 months. There is a good chance that whatever is being proposed, even if exaggerated in the moment, will likely happen.
We see this occurring on a regular basis.
AI Scientist Gets Its First Peer Reviewed Paper
Sakana is now to the point where it claims to have created its first peer reviewed paper. This means the paper was developed completely by AI, followed by the traditional submission process to get reviewed.
This is another milestone in the AI evolution.
Some are contesting the exact level of achievement. There were some noted errors such as misidentifying a year in another work cited. This could not change the outcome of the situation a great deal but does show that embarrassing errors can still occur.
The company said that it used an AI system called The AI Scientist-v2 to generate a paper that Sakana then submitted to a workshop at ICLR, a long-running and reputable AI conference. Sakana claims that the workshop’s organizers, as well as ICLR’s leadership, had agreed to work with the company to conduct an experiment to double-blind review AI-generated manuscripts.
It appears three papers were submitted and one was accepted using a double blind system.
As always, it is easy to get caught up in the details. If a misstating of a date is the most crucial error, then I think the system is doing well. Of course, we cannot say it is 100% accurate and human interaction would be required.
That said, as always, we have to remember this is the worst the model will be. It will only get better from here.
It is helpful this is open source so anyone can download it and advance the system.
AI Penetrating All Industries
Sakana is not quite up to the level of co-scientist. Nor is the system ready for AI scientists to emerge. The review did not enter a post-review phase, something that was agreed upon before hand.
In other words, the AI model is not yet published in this regard.
Here again, simply wait 6-12 months and the story will be completely different. This holds true for any industry. Software is being developed to interact with LLMs, creating different results from what humans do. This will be enhanced as these models keep advancing.
With the AI scientist, in spite of the shortcomings, it did bring new knowledge to the table. While the paper didn't quality to meet the highest of standards, the team did assert that it provided areas for further research.
Here we are seeing where a lot of breakthroughs will happen. As AI models get better, they can provide preliminary research in all kinds of fields. Even if further AI inquiry is stalled, humans can run with it.
This could save years worth of human research time.
These specialized systems will likely be commonplace by the end of the year.
Posted Using INLEO
The rewards earned on this comment will go directly to the people( @x-rain ) sharing the post on Reddit as long as they are registered with @poshtoken. Sign up at https://hiveposh.com. Otherwise, rewards go to the author of the blog post.
Sure, it made some errors, but so do human scientists. We give it a year, and it'll probably be correcting them too. I love the speed of our advancements in all tech
Yeah but humans can at least edit it properly (for the most part).
absolutely Mr Task. This thing should be considered a big ass revolution
I hope that AI will contribute to science, but we need to be careful so we dont rely on it to much. We still need sharp human minds on this planet
Science is the field where I have the most expectations for AI. It is because I think "the knowledge is already there", but simply there are no human scientists that cover different fields of knowledge to notice connections. Over the years and centuries sciences became more and more specialized and so did the scientists. It is no longer humanely possible to be an expert in multiple fields like Leonardo da Vinci.
Let's say there is a common plumbing problem that workers learn to work around as they gain experience. Let's also say the problem is similar to some unsolved question in the field of applied fluid dynamics and the solution exists based on quantum physics. How many expert plumbers that also know those two fields of science? Probably not many. But AI could have knowledge in all those areas and come up with connections naturally through emergence.
This is quite concerning and perhaps fears of getting AGI soon are near.
AI will continue to get better as it is being fed a lot of informations
Now it is really looking like daily, we are having an AI world carving in. It is becoming a reality as the day goes by