In surprising messages, the co -founder and general director of the Openai Altman himself has a blog. And among browsing classic topics similar to blogs, such as “What and Wisand Your Fall Me” and “Strength of being misunderstood”, he recently published Three AGA observations (Artificial general intelligence) and its potential applications of the human race.
“Economic growth ahead of us looks amazing and we can now imagine the world in which we treat all diseases, we have much more time to enjoy our families and fully realize our creative potential,” says Altman.
“In a decade, maybe everyone on Earth will be able to achieve more than the most influential person.”
Well, it sounds wonderful, right? Who does not like the promise in the style of Star Trek, utopian, after prudent, after madness, in which our efforts are supported by our private genius. To justify this thinking, three Altman’s observations are therefore:
1. The intelligence of the AI model is roughly equal to the resource diary used for training and conducting. “These resources are mainly training of calculations, data and inference. It looks like you can spend any amounts and get continuous and predictable profits; Calculating regulations that provide that they are accurate in many rows of size. “
2. The cost of using a given level of artificial intelligence drops about 10 times every 12 months, and lower prices lead to much more operate. “You can see it at the cost of the GPT-4 token at the beginning of 2023 to GPT-4O in mid-2012, where the price for the token fell about 150x at that time. Moore law changed the world to 2x every 18 months; This is incredibly stronger. “
3. The socio-economic value of a linearly growing intelligence is very exponent. “The consequence of this is that we see no reason to exponially increase investments to stop in the near future.”
This last point seems particularly vital, considering that Opeli was previously reported that it burns billions of dollars in training and modeling costs. The cooperating investment was a key contemporary AI boom, and the release of the Chinese startup Deepseeka R1 (apparently trained for a fraction of the costs of existing efforts) recently shocked the trust of investors in the AI industry dominated by the US.
No wonder that Altman emphasizes its meaning here. However, at first it looks like Altman has no illusions from the mainstream of the artificial intelligence market (if there is such a thing) resignation from the costs of training, hardware requirements and “any amounts of money” to continue to gain land in AI development, at least when it comes to About Aga.
Despite this, according to Altman, at least he becomes cheaper in time. Which, if his forecasts about the future of AI agents come true, are necessary to enable our AI-CO-Vorker Hellsc … I mean future working methods.
“Imagine the case of software engineering agent … Imagine it as a real virtual colleague.
“The world will not change at the same time; It never happens. Life will act mainly in the miniature term, and people in 2025 will spend time mainly in the same way as in 2024. We will still fall in love, create families, fight on the Internet, wandering in nature, etc.
“But the future will come to us in a way that cannot be ignored, and long -term changes in our society and the economy will be huge.”
Candy. I am content to hear that in Altman’s eyes I will continue to fight online and wander in nature this year. But the assistants from Agi-observed are coming, says Honcho Opeli, and taking into account the previous trends he emphasizes here, they seem that they are coming quite quickly (of course, provided that the tap flows with money).
“The agency, defectiveness and determination will probably be extremely valuable,” continues Altman. “The correct decisions about what to do and thinking on how to move in a constantly changing world will have great value; Resistance and adaptability will be helpful skills to cultivate. “
“Agi will be the largest lever of human will and will allow individual people to have a greater influence than ever before, no less.”
It seems that a lot of bullish thinking is happening here. I would like to raise my hand and say that I am not very interested in the idea that AI’s colleague writes my articles for me, but if they could “endure intellectual ability” to reliably lose my inbox without sending vital messages to a spam folder that would be great.
I don’t want Aga Mozart, more competent Jeves. Despite this, as Altman has it, everything sounds suspicious and pink for our innovative future:
“At the moment there are many talents without resources to fully express ourselves, and if we change it, the resulting creative performance of the world will lead to huge benefits for all of us.”
I hope anyway.
