AI

DeepSeek’s R1: Helpful reminder

As a college educator and a former IT industry veteran, I found the hype around the Chinese Deepseek R1 model to be a useful reminder, reminding three things.

The first is that generating AI is no longer just processing a lot of content to produce relevant responses to prompts. This is also related to cognitive reasoning (“R” in R1).

The hope of the Great Language Model of Reasoning (LLM) is that a lot of knowledge retrieval and cognitive processing capabilities—one that were exclusive areas of the brain with supercomputers—now almost all in the hands of everyone. Thanks to the advancement of the next-generation efficiency-enhancing technology, there are small enough models to run on traditional laptops that can support multiple smart agents that can perform complex, interactive tasks autonomously.

Second, the most important thing about the generated AI revolution is innovation and creativity – it is not only an arms race against the most powerful hardware, the size of the training dataset, or the number of model parameters. The successful adoption of these technologies will not be the billion-dollar model trained by large technology companies, but by investing in human capital to invest in countries and organizations ready for the new wave.

Third, at the last point, the United States does not seem to be beneficial to the huge changes that our economy and society bring. I will give two examples: higher education and American companies.

High-rise

In most higher education institutions, the first major decision for undergraduates is to decide whether to pursue a Bachelor of Arts (BA) degree, which is related to a broader, more interdisciplinary Bachelor of Education or Science (BS) degree. More Focus on developing skills and practical experience in specific areas.

In the AI ​​era, this is a desperate dichotomy, as both groups of disciplines become crucial in the workplace.

The truth is that most first-year students have no knowledge or insights about the relative strengths and weaknesses of their abilities, talents, skills and abilities. However, most first-year graders have to declare a major, which is for only a small number of people (for better or worse) know (or at least think they know) what areas they want to pursue: Engineering, Engineering, Science, Medicine, Law wait

We need to take several different, career, broader interdisciplinary approaches to achieving higher education, acknowledging that the first full-time job of a college graduate may not be related to the degree or degree they receive. Their college experience will represent only the first stage of lifelong learning, a role that we cannot even imagine now, continuous learning – high skills, certification, reshaping, career transition.

Similarly, as educators, we need to develop new strategies to deal with AI plagiarism and lead to the danger of chatbots becoming intellectual shortcuts or “cognitive offloading” or the tendency to rely on external tools rather than developing internal functions.

In an era where knowledge and understanding are separated, there are too many temptations to just prompt AI instant answers or solutions instead of falling to understand concepts or solve problems.

company

Most companies also don’t seem to realize the organizational meaning of these new technologies.

The current IT role and structure reflects the organizational requirements of the previous digital revolution. These features are derived from the professional expertise required Human Use and interact computer – Programming, data engineering, computer architecture, network management, information security, etc.

By contrast, generated AI (and its entire field of natural language processing before it) involves design and training computer Interact with it Human.

As a result, grade and document employees are inventing excellent (and sometimes dangerous) ways to use these technologies. Organizations are working to propose viable policies, procedures and controls to maximize potential productivity benefits while minimizing risks.

A key issue is that in most companies, data science expertise tends to be concentrated in IT departments, most of whom still operate in secret actions that are organizationally and functionally isolated from core business units in their own mysterious language and practices. I believe the upcoming productivity revolution requires new types of organizational roles and structures in which data expertise is not isolated in professional functions but is interconnected with almost every aspect of operation.

And there are data challenges. In most organizations, adopting AI is about custom LLMs using proprietary data to perform professional use cases. Although data users within the business scope require complete accuracy, clean and good data, the individual owners of the data do not have budget, financial incentives or organizational authority to ensure this level of quality and transparency.

As a result, internal datasets cannot be discovered/hosted throughout the enterprise. Usually, different types of data are stored in different places. In response to requests from business users, it provides different data views, makes different copies of the data (and copies) and creates exposure and abstraction of the data for a variety of different reasons… At this point, no one knows which versions are Old, incomplete, repeated, inaccurate or their background.

in conclusion

Generated AI has the potential to change all forms of knowledge work. Essentially, this technology is about democratizing expertise (good vs. bad) – experts like coders, videographers, illustrators, writers, editors, and almost any type of knowledge worker or “expert”. Humans have never dealt with a technology that rivals their own cognitive processing and reasoning abilities – just their physical strength, endurance, agility, and the ability to invent and process large amounts of data.

This exciting new productivity revolution requires new skills, capabilities and organizational structures where data expertise is integral to nearly every type of business process.

Ironically, as machines gain greater analytical capabilities, the status and value of employees in an organizational hierarchy may no longer be the function of expertise, experience and certificates, and more of its creative, multidisciplinary and interpersonal skills. .

Now is the time to develop and invest in these features.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button