If you want to use artificial intelligence in business then start learning linear algebra without know how to interact directly with models you will lack the understanding of why the engineers already knew the limitations.
Tatyana Fomina, a recruiter, shares how she built a custom GPT to automate text-based tasks, optimize job ads, and train colleagues, highlighting the importance of customization and overcoming challenges with large language models. Key moments: 00:10 The video discusses the importance of using a custom GPT in recruitment to increase productivity and efficiency. The speaker shares insights on building a custom GPT and its impact on optimizing speed and quality in candidate communication. -The speaker highlights the competitive nature of recruitment and the challenges faced in the talent market in Europe, emphasizing the need for speed in candidate acquisition. -The discussion shifts to the use of AI in recruitment to streamline communication, optimize touchpoints with candidates, and address the overwhelming volume of conversations recruiters manage. 06:05 Custom GPTs were challenging to optimize prompting, leading to the development of a simpler approach for generating responses, emphasizing the importance of adapting to evolving technology in corporate settings. -The speaker highlights the necessity of embracing AI as a crucial aspect of their work in a corporate and technological environment, emphasizing the inevitability of technology's presence and the need to adapt to it. -The discussion shifts to the international hiring process, focusing on the ease of relocating talent globally due to skill shortages in specific regions, highlighting the role of agencies in facilitating such relocations. -The speaker emphasizes the importance of learning from experienced individuals, utilizing resources like Reddit for guidance on working with custom GPTs efficiently, and avoiding unnecessary trial and error in the learning process. 12:08 Using GPT for writing job ads can save time and improve quality by providing templates based on preferred examples, allowing focus on creativity and high-impact content, ultimately enhancing productivity and scalability. -Challenges of writing job ads manually and the benefits of using GPT for creating repeatable and high-quality content, especially for passive candidates and scalable tasks. -The process of utilizing GPT to generate job ad templates by feeding it preferred examples, refining the output for language and regional specifics, and the significant impact on productivity and content quality. 18:09 Using GPT models to generate job ads can be highly efficient and effective by providing a structured template and asking specific questions to ensure accurate and relevant information is included. -Customizing GPT prompts with specific questions can enhance the quality of generated content and streamline the job ad creation process. -Implementing a list of banned words in GPT models can improve the quality of generated text by avoiding biased or irrelevant language, leading to sharper and more effective job ads. 24:12 Using the 'configure chat' feature can lead to unintended changes in instructions, emphasizing the need for detailed prompts to guide GPT responses effectively. -The importance of using 'configure chat' feature cautiously to avoid unintended changes in instructions and the need for minimal interaction after initial setup. -The significance of providing detailed prompts to GPT to ensure accurate and consistent responses, highlighting the impact of over-explaining rather than under-explaining. -The role of individuals with expertise in literature and psychology in training GPT, showcasing the nuanced understanding required for effective communication with language models. 30:15 Understanding the impact of removing one word from a prompt on large language models can significantly alter outcomes, highlighting the importance of word choice and its influence on predictions and results. -Importance of experimentation and word choice in prompts for large language models, emphasizing the need for strategic input to achieve desired outcomes. -The role of TH-cam as a valuable resource for information on large language models, showcasing the significance of understanding the processes behind these models for enhanced user experience. -Challenges and limitations of large language models in interpreting common sense and context, leading to unpredictable responses and the focus on grammatical correctness over semantic meaning. 36:17 The speaker explains how they had to teach their colleagues to work with custom GPTs, emphasizing the need to adjust prompts for accurate predictions and the challenge of changing established work habits. -Challenges in teaching colleagues to use custom GPTs and the importance of adjusting prompts for accurate predictions. -The difficulty of changing established work habits when introducing new tools and the need for patience and guidance. -The speaker's strategy of selecting early adopters and overcoming initial resistance to new technology in the workplace. 42:23 Using GPT to streamline tasks can save time and increase efficiency significantly, leading to unexpected positive outcomes and the ability to train individuals in various industries with GPT. -Benefits of using GPT for task optimization and efficiency. GPT can handle mundane work, save time, and improve quality of output. -Surprising outcomes and the potential of training individuals with GPT. Unexpected positive results and the ability to train people in different industries using GPT. -Customizing GPT for specific needs. The importance of building custom GPT models for niche professions to enhance productivity and effectiveness. 48:29 To become an AI expert, stay curious, test things, make mistakes, and learn from them. Being an expert is about going the extra mile, not necessarily having a formal title or background in computer science. -The importance of making mistakes and learning from them to gain expertise in AI. Expertise comes from experience and willingness to go the extra mile. -The misconception of needing a formal background in computer science to understand AI. Expertise is more about practical experience and learning from mistakes. -Encouragement to take initiative in learning AI early to gain reputation, experience, and trust in the industry, rather than waiting for expensive courses. Generated by sider.ai
![[1hr Talk] Intro to Large Language Models](http://i.ytimg.com/vi/zjkBMFhNj_g/mqdefault.jpg)
![[Full Episode] The Restaurant War Thailand ศึกพ่อค้าซ่าแม่ค้าแซ่บ Episode 2 | 29 ก.ย. 67](http://i.ytimg.com/vi/r3lfLNH6Ke4/mqdefault.jpg)
Won't AI eventually replace all od yhe tasks she is doing?
If you want to use artificial intelligence in business
then start learning linear algebra
without know how to interact directly with models
you will lack the understanding of why the engineers already knew the limitations.
That’s one approach but impractical for most. I’ve studied all of that and it’s a lot of work. Better to have someone consult you practically.
@@MarcelSamyn yes, but the consultant will just hire a third consultation party.
Tatyana Fomina, a recruiter, shares how she built a custom GPT to automate text-based tasks, optimize job ads, and train colleagues, highlighting the importance of customization and overcoming challenges with large language models.
Key moments:
00:10 The video discusses the importance of using a custom GPT in recruitment to increase productivity and efficiency. The speaker shares insights on building a custom GPT and its impact on optimizing speed and quality in candidate communication.
-The speaker highlights the competitive nature of recruitment and the challenges faced in the talent market in Europe, emphasizing the need for speed in candidate acquisition.
-The discussion shifts to the use of AI in recruitment to streamline communication, optimize touchpoints with candidates, and address the overwhelming volume of conversations recruiters manage.
06:05 Custom GPTs were challenging to optimize prompting, leading to the development of a simpler approach for generating responses, emphasizing the importance of adapting to evolving technology in corporate settings.
-The speaker highlights the necessity of embracing AI as a crucial aspect of their work in a corporate and technological environment, emphasizing the inevitability of technology's presence and the need to adapt to it.
-The discussion shifts to the international hiring process, focusing on the ease of relocating talent globally due to skill shortages in specific regions, highlighting the role of agencies in facilitating such relocations.
-The speaker emphasizes the importance of learning from experienced individuals, utilizing resources like Reddit for guidance on working with custom GPTs efficiently, and avoiding unnecessary trial and error in the learning process.
12:08 Using GPT for writing job ads can save time and improve quality by providing templates based on preferred examples, allowing focus on creativity and high-impact content, ultimately enhancing productivity and scalability.
-Challenges of writing job ads manually and the benefits of using GPT for creating repeatable and high-quality content, especially for passive candidates and scalable tasks.
-The process of utilizing GPT to generate job ad templates by feeding it preferred examples, refining the output for language and regional specifics, and the significant impact on productivity and content quality.
18:09 Using GPT models to generate job ads can be highly efficient and effective by providing a structured template and asking specific questions to ensure accurate and relevant information is included.
-Customizing GPT prompts with specific questions can enhance the quality of generated content and streamline the job ad creation process.
-Implementing a list of banned words in GPT models can improve the quality of generated text by avoiding biased or irrelevant language, leading to sharper and more effective job ads.
24:12 Using the 'configure chat' feature can lead to unintended changes in instructions, emphasizing the need for detailed prompts to guide GPT responses effectively.
-The importance of using 'configure chat' feature cautiously to avoid unintended changes in instructions and the need for minimal interaction after initial setup.
-The significance of providing detailed prompts to GPT to ensure accurate and consistent responses, highlighting the impact of over-explaining rather than under-explaining.
-The role of individuals with expertise in literature and psychology in training GPT, showcasing the nuanced understanding required for effective communication with language models.
30:15 Understanding the impact of removing one word from a prompt on large language models can significantly alter outcomes, highlighting the importance of word choice and its influence on predictions and results.
-Importance of experimentation and word choice in prompts for large language models, emphasizing the need for strategic input to achieve desired outcomes.
-The role of TH-cam as a valuable resource for information on large language models, showcasing the significance of understanding the processes behind these models for enhanced user experience.
-Challenges and limitations of large language models in interpreting common sense and context, leading to unpredictable responses and the focus on grammatical correctness over semantic meaning.
36:17 The speaker explains how they had to teach their colleagues to work with custom GPTs, emphasizing the need to adjust prompts for accurate predictions and the challenge of changing established work habits.
-Challenges in teaching colleagues to use custom GPTs and the importance of adjusting prompts for accurate predictions.
-The difficulty of changing established work habits when introducing new tools and the need for patience and guidance.
-The speaker's strategy of selecting early adopters and overcoming initial resistance to new technology in the workplace.
42:23 Using GPT to streamline tasks can save time and increase efficiency significantly, leading to unexpected positive outcomes and the ability to train individuals in various industries with GPT.
-Benefits of using GPT for task optimization and efficiency. GPT can handle mundane work, save time, and improve quality of output.
-Surprising outcomes and the potential of training individuals with GPT. Unexpected positive results and the ability to train people in different industries using GPT.
-Customizing GPT for specific needs. The importance of building custom GPT models for niche professions to enhance productivity and effectiveness.
48:29 To become an AI expert, stay curious, test things, make mistakes, and learn from them. Being an expert is about going the extra mile, not necessarily having a formal title or background in computer science.
-The importance of making mistakes and learning from them to gain expertise in AI. Expertise comes from experience and willingness to go the extra mile.
-The misconception of needing a formal background in computer science to understand AI. Expertise is more about practical experience and learning from mistakes.
-Encouragement to take initiative in learning AI early to gain reputation, experience, and trust in the industry, rather than waiting for expensive courses.
Generated by sider.ai