the paper research | the latest, biggest and strongest things every day, are you annoyed by ai?
2024-09-12
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
ai's iphone is not an iphone at all times.
we have finally waited for apple to release its first ai phone in history. after people’s expectations were raised in advance, they found that there was no breakthrough innovation, and users outside of some english-speaking countries were temporarily unable to experience apple’s smart phones.
as early as a year and a half ago, huang renxun said: a brand-new computing platform has been invented, and the iphone moment of ai has begun. however, this computing platform refers to chatgpt, and there is no second equally powerful competitor so far.
is it so difficult for apple to develop ai? is the competition landscape in the field of ai a competition among many players or an oligopoly?
who is being given the expectations for the development of ai?
paike ai unboxing and the paper news align lab jointly released the "artificial intelligence new media content communication power insight report", based on 3,517 reports on ai on the paike platform in the first half of 2024the analysis found that the first place is almost a matter of course, but the second place result is worth careful consideration.
a research report from ubs group shows that it took openai's chatgpt intelligent chatbot two months to reach 100 million monthly active users, while tiktok took nine months to achieve this goal. the well-known silicon valley investment institution a16z has published three ai application top 100 lists, and chatgpt has always ranked first. therefore, among the top 10 most popular companies, openai's popularity index is nearly twice that of the second place.
apple is the second most mentioned company, not because of its success in ai, but because its ai progress has made people too anxious. although the company, which has always been known for its technological innovation, still has a high tone in its publicity, it has now become obviously weak and lackluster. "apple saves itself", "really anxious", "internal and external troubles", "dilemma in the ai era"... in the ai era, even technology giants have to compete.
alibaba, baidu, and bytedance are also among the top 10 companies mentioned most in the headlines., tencent followed closely behind, ranking 11th. these traditional internet giants are working hard to roll out ai, and no one wants to fall behind in the ai era.
there are also some young ai companies that, despite lacking scale and first-mover advantage, have also made some eye-catching products, and even the products are more famous than the companies. the most representative one is "kimi". in the domestic ai product list for august released by the "ai product list", the first and second places are bytedance's "doubao" and baidu's "wenxin yiyan", and the third place "kimi" comes from the dark side of the moon, a company that was only established in 2023. in the titles of the articles we analyzed, "kimi" was mentioned 5 times more than the dark side of the moon. in addition, minimax, zhipu and the dark side of the moon are all new "ai four little dragons" in the era of generative ai, and they also have products that are in the top 10 of the list.
but the breakthrough of young companies is also made with real money. according to media reports, "kimi" is associated with almost all the terms that users may search for "ai" on bilibili, and the cpa (cost per action, the average cost of user conversion) quoted by marketing is as high as 30 yuan, which makes other ai companies that invest in bilibili feel pressured. does the high cost of acquiring customers mean that burning money is the bottom line? will ai competition follow the old path of traffic competition in the mobile internet era? how long can small companies last in the face of the great waves?
as for ai-related people, we found thatthe person the creator mentioned most in the title was not openai's ceo altman, but musk, who always confronts him.
over the past year, musk and altman have frequently exchanged insults, suggesting that openai be renamed "closedai," and accusing openai of breach of contract and filing a lawsuit. is musk too idle or too idealistic?
in response to musk's accusations, openai claimed that if the case begins pre-trial fact-finding and information sharing, musk will use the lawsuit to obtain openai's "proprietary records and technology." some people also pointed out that openai also wants to use this to destroy musk's personality.
in fact, ai has long been a battleground for stakeholders in various fields, and musk already has a lot of ai puzzle pieces in his hands: xai, an artificial intelligence company founded last year, the large model grok-2, the brain-computer interaction neuralink, the supercomputer dojo, and the robotaxi that tesla will release in october, etc.
the battle in the business field never lacks attention. and there are also many developments in the academic giants in the field of ai. in april, the team co-led by "ai godmother" fei-fei li released the "2024 artificial intelligence index report". on the paike platform, a related report ranked in the top 30 in terms of clicks. since then, the news that fei-fei li started a space intelligence business has been widely reported. there are even more reports about turing award winner and meta chief scientist yann lecun - from publishing research papers, being selected into the acm fellow list of the american computer association, to "severely criticizing" sora on the social platform x, and starting a "war of words" with musk...
ai is improving every day, but do ordinary people still care?
those who pay attention to ai will certainly recall that in mid-february, before the chinese new year, the wensheng video model sora launched by openai continued to dominate social media. voices such as the film and television industry has changed and reality no longer exists have been heard. in addition, new large models and products such as stable diffusion 3 and google's gemma also appeared in february. meta announced its "strongest financial report in history" in early february and announced its first dividend distribution in history.
but ai's highlight moment in the first half of the year ended in february. in the four months since then, no other technology product has been as popular as sora.creators' enthusiasm for ai reporting is still rising, but readers' enthusiasm has already "ebbled"——the overall click volume has declined significantly, and the median click volume index in june is almost half of that in february.
although there have been no new disruptive or revolutionary changes in technology and products, ai news titles still try to attract readers with exaggerated words such as "biggest", "strongest" and "latest". "we can't stand the current ai self-media style of subverting every three days and revolutionizing every two days." the host joked in a podcast program this year that talked about "active actors" in the ai era.
will readers get tired of seeing too many exaggerated words? we selected 12 high-frequency words with strong exaggerations from the titles of 3,517 related reports on the paike platform in the first half of the year.
it can be seen that readers may still be attracted by expressions such as "strongest", "hottest", "biggest", "shocking", and "heavyweight", but the effects have mostly weakened over time. although "latest" appears most frequently in titles, it does not play a big role in increasing article clicks. the median click index of related articles is even lower than that of other articles (titles that do not contain these 12 high-frequency words). after all, the essence of news is the "latest" events, trends, and phenomena.
ai's traffic code: funny, offensive, and job-stealing
at this stage, the possible state of most people facing ai is: they can chat a little, but not in depth; they will care a little, but not much.
the artificial intelligence academic and industry communities publish new research papers every day, but ordinary people are more concerned about which new ai products and tools are easy to use; ai has penetrated into fields such as medicine and biology, and has already brought about some scientific breakthroughs, but the question "will ai make programmers unemployed?" makes ai look more capable and more threatening.
we used k-means topic clustering analysis to classify 3,517 articles into 23 major topic categories (including "others"), and ranked the readers' preference for each topic based on the median number of clicks on articles under different topics.
we found that:the topics that creators write about most are mostly related to the latest developments in the field of ai, covering academic discoveries, technological research, company and product trends, etc.after all, one day in ai is like one year in the human world. in the past six months, various ai products have emerged in an endless stream and have been released intensively, and academic research has also been quite active. according to the ai survey report released by eto, a data analysis platform of georgetown university in the united states in may this year, a total of 1.255 million english technical papers related to ai were published worldwide from 2017 to 2022. after chatgpt and sora became popular, related research will inevitably grow rapidly again.
however, compared with foreign ai companies, products and trends, readers are more concerned about the situation in china.compared with ai itself, readers are more concerned about the industries and professions affected by aithe two most clicked topics - automobiles and e-commerce - seem to have nothing to do with ai, but they are closely related to the lives of ordinary people. because of the influence of ai, many unknowns and topics have emerged.
the creators’ enthusiastic pursuit of ai dynamics and the readers’ calm reaction are in stark contrast. but looking back at the technological craze since the millennium, the “technological cooling-off period” after the craze is often accompanied by a transition from irrational prosperity to rational development. for new technologies, the public’s relative calmness may be a good thing.
in an interview with the creator of "jia zi guang nian", the famous technology scholar kevin kelly expressed that the real concern about ai is "weaponization" and "there are some problems in the field of ai that we don't know how to solve, and these problems may cause us trouble in the future."
if such concerns are still long-term concerns, ordinary people already have their own immediate worries.according to the "artificial intelligence public attitude survey report" published by the paper in march this year, 54.8% of respondents would feel uneasy about the decision-making power of ai in the medical and health fields such as medical diagnosis, and 41.8% of respondents would worry about the application of ai in the field of personnel management, such as hiring and firing employees. in addition, 50.5% of respondents "somewhat agree" that ai is likely to get out of control, and only 14.4% think that ai will not get out of control.
in the first half of this year, technology has also caused some naked real-life problems: li yizhou's ai course was exposed as fake; ai face-changing fraud cases occurred frequently; ai paintings squeezed out the living space of painters... in the first half of this year, ai-related topics also frequently appeared on hot searches. among them,the "opportunities" and "risks" brought by technology go hand in hand.
in addition to its impact on humans, ai can also arouse the interest of ordinary people simply by virtue of its own charm. only ai that can "make things happen" can be "lovable". among the top three articles with the highest click-through rate on the paike platform, two are related to "ai making things happen". in the track of ai creators, a genre has emerged that specializes in testing or mining various "ghostly" videos made with ai. now, "reverse making things happen" where humans imitate ai has become popular.
in addition to musk, altman and others, which other people's stories about ai are people's concern? what trends are worth paying attention to behind the social phenomena caused by ai? which creators are the most productive in the field of ai? which creators are not in the field of science and technology, but their reports on ai have also become popular?
for more details, please refer to the article “who still cares about ai when the craze is cooling down? ——2024 ai new media content communication power insight report” based on the answer was found in the analysis of 429 creators and 3517 ai-related reports. scan the qr code on the poster to view the original text.
the paper reporter zhang lingyuan, wang yasai, special author chen liwei, intern tian xiuqi, lin xinyi, zhu yidan
(this article is from the paper. for more original information, please download the "the paper" app)
