科技周会 | 人工智能/AI/GPT

Host: Tracy (@tracyyxchen), Roger (@rogerzhufilm)

This podcast episode explores the recent funding trends in AI startups, the challenges faced by smaller startups when competing against big players like Microsoft and Meta, the concept of prompt engineering and its future, the implications of AGI development, the definition of AI and its performance level, the potential of AI in content creation, the concept of machine unlearning in large language models, and the use of AI in PDFs. The conversation also touches on the collaboration between Meta and RayBan, the developments of big tech companies in AI, and the difficulty of finding tech products that live up to their promises.

Takeaways

• AI startups have been raising billions of dollars in funding, but concerns about a potential bubble arise due to the lack of revenue in many startups.
• Smaller startups face challenges when competing against big players like Microsoft and Meta, leading to dissolution for some teams.
• Prompt engineering may become less relevant as large language models like OpenAI simplify the prompts process, but there will still be special use cases for prompt engineering.
• The development of AGI and its implications for humans' existence and prompt engineering are topics of ongoing debate.
• The definition of AI and the desired performance level of AI are discussed, with predictions of fully functioning models fulfilling 99% of desired capabilities in the next two years.
• AI has the potential to revolutionize content creation, with advancements expected in generating scripts, 3D models, and videos.
• Machine unlearning, the ability to make models forget specific data, is a topic of interest for the future as the need to forget confidential or harmful content arises.
• AI features in software like ReadCube and Adobe Acrobat enhance the use of AI in PDFs, allowing for Q&A interactions and understanding complex documents.
• Tracking the AI advancements of big tech companies such as Meta, Apple, and Google is important, and finding tech products that deliver on their promises can be challenging.

Outlines

The Funding Landscape for AI Startups: Evaluations Soaring to New Heights
00:00Startups and AI Trends: Dissolving Startups, Impressive AI Demos, and Learning Prompts
03:33The Future of Prompt Engineering and the Evolution of Large Language Models
07:17Thoughts on Prompt Engineering and AGI Arrival
10:59Progress towards Advanced AI and the Future of Content Creation
13:46The Future of AI Content Creation and Machine Unlearning
16:25Machine Unlearning in 2024: Can We Make Models Forget?
18:21Unlearning Techniques and AI Companions
21:12AI and Technology Innovations Discussed, Rabbit R1 Mentioned
24:32
前往小宇宙评论区与主播互动

This episode is sponsored by: gratitudejournal.ai

Try it here: app.gratitudejournal.ai

The magic power of keeping a gratitude journal. Join by contact: [email protected]

前往小宇宙评论区与主播互动

This podcast episode explores significant developments in the electric vehicle market, including Xiaomi's entry with an affordable and Porsche-inspired car, Apple's decision to halt their EV project, and Tesla's challenges with device integration. The discussion also touches on luxury electric cars like the Porsche Taycan and the popularity of flexible screen technology, with a focus on Samsung's foldable phones. Additionally, the episode highlights the Norwegian Sovereign Wealth Fund's massive size and impact on the country's economy and discusses Alibaba's strategy to refocus on its core e-commerce business.

Takeaways

• Xiaomi's affordable electric vehicle, resembling luxury brand Porsche, is generating buzz in the market.
• Apple's decision to halt their EV project raises questions about safety concerns and user experience.
• Porsche Taycan is considered a symbol of luxury and status in the electric vehicle industry.
• Flexible screen technology, exemplified by Samsung's foldable phones, is gaining popularity.
• Norwegian Sovereign Wealth Fund's management of a significant amount of money showcases challenges faced by other countries.
• Alibaba aims to refocus on its core e-commerce business to regain market share.

Outlines

Xiaomi Enters the Electric Vehicle Market, Drawing Comparison with Porsche
00:00Apple Stops EV Project, Xiaomi's Impressive Launch, and Tesla's Integration Struggles
02:28Porsche Taycan and the Downturn of Flexible Screens
06:40Flexibility of Screens and the Norwegian Sovereign Wealth Fund
09:50Alibaba's New Direction and Focus on Market Share
12:22
前往小宇宙评论区与主播互动

Host: Tracy (@tracyyxchen), Roger (@rogerzhufilm)

This podcast episode explores various topics related to technology and entrepreneurship. The speakers discuss the importance of fast response times in hardware products, such as the Rabbit R1 AI hardware and Groq, and the potential benefits of augmented reality (AR) devices in productivity. They also analyze the pricing strategies of AI models, including the uniform pricing of $20 per month and the visual aesthetics of Cloud. Additionally, the challenges faced by ride-sharing apps in attracting both drivers and passengers are explored, with a focus on Didi's successful implementation. The episode concludes with a discussion on the consequences of stealing internal documents from tech companies and the importance of ethical conduct.

Takeaways

• Fast response times are crucial for hardware products, such as the Rabbit R1 AI hardware and Groq, to provide a better user experience.
• Augmented reality (AR) devices have potential in improving productivity, but their high cost may limit their adoption in schools.
• The pricing strategy of AI models, like the uniform price of $20 per month, may not accurately reflect their capabilities and performance.
• Cloud's UI design, resembling a traditional newspaper, may appeal to users who prefer a human-centered interface.
• Attracting both drivers and passengers is a challenge for ride-sharing apps, and strategies are needed to break the "deadlock" situation.
• The importance of ethical conduct in business is emphasized, and the serious legal consequences of stealing internal documents from tech companies are highlighted.

前往小宇宙评论区与主播互动

Host: Tracy (@tracyyxchen), Roger (@rogerzhufilm)

This podcast episode explores the rise of Perplexity AI, a chatbot that has garnered attention for its impressive performance. The speakers question Perplexity AI's valuation and compare it to ChatGPT, discussing the potential impact of the network effect and different company focuses. The episode also examines the success of Rivian, an electric vehicle company that is competing with larger companies like Tesla, highlighting the potential benefits of focusing on a specific niche. The speakers conclude by discussing the government's potential support for Rivian and the challenges faced by Chinese EV companies.

Takeaways

• Perplexity AI's success may be driven by the network effect, similar to social media recommendations.
• Rivian's focus on a specific niche in the electric vehicle market allows it to compete with larger companies like Tesla.
• The government may support Rivian to prevent Tesla from becoming the only major EV company in the US.
• Chinese EV companies face challenges due to tariffs.
• AI-generated content may impact Google's search algorithm.

Outlines

PerplexityAI: A Unicorn with a Browser Plugin
00:00Perplexity AI vs. ChatGPT: Network Effect and Focus Drive Success
02:17Rivian: A Smaller Company Competing with the Giants
03:54Rivian's Government Support and the Future of EV Companies in the US
05:06

Q & A

Q: How do you think about Perplexity AI?

A: I feel ChatGPT also has the same capability. It has a browser plugin and it also starts with a chat.

Q: What is the difference between perplexity AI and OpenAI?

A: Perplexity AI focuses on search and optimizes the user experience, while OpenAI focuses on being a platform and wants others to use its capabilities to build other things.

Q: What is the relationship between Rivian and Tesla?

A: Rivian is a smaller company that is more focusing on one specific area and trying to be the best in that, while Tesla is a larger, more comprehensive company that's maybe 90% good at everything.

Q: What is the reason why the government wants to keep Rivian alive?

A: Because they don't want Tesla to be the only pure EV company in the US, which could lead to antitrust lawsuits.

前往小宇宙评论区与主播互动

This podcast episode explores the potential and challenges of using AI tools like Sora and Apple Vision Pro for video creation. The speakers highlight the advantages of Sora in generating realistic videos with physics-based movements, and discuss the potential cost limitations of the tool. They speculate about the use of AI-generated videos for prototyping, and compare the rapid progress in the AI world to the development of mobile internet technology.

Takeaways

• Sora has the potential to revolutionize video creation with realistic and physics-based video generation.
• AI-generated videos can serve as prototypes or drafts for real-world video production.
• Mid-Journey's success in image generation suggests the potential for disrupting stock photo websites.
• The gap between China and the US in AI technology and human power may be smaller than previously thought, as evidenced by the dual roles of Chinese AI professors.

Outlines

Sora: A Revolutionary AI Tool for Creating Realistic Videos
00:00Sora: A Potential Game-Changer in Video Production
02:49AI Video Generation: Sora and Apple Vision Pro, MidJourney's Hardware and Collaboration with Xtrader
06:21Mid-Journey and China's AI Investment Landscape
09:22Chinese AI Professors' Dual Roles: Balancing Research and Startups
12:26

Q & A

Q: What are the advantages of Sora over Pica Labs?

A: Sora can support as long as 60 seconds, and the details are so real. And it seems like every movement of those objects, they follow the physics in real life.

Q: What kind of business are you going to use as generative videos?

A: One example is that like, you know, before this, when people have to shoot like a movie or like a well-made video, first they need to make like a a screenplay so that's like drawings of static frames so now with Sora you can probably generate like a pseudo like draft of the video rather than just have like static like hand drawn images so that improves like the drafting stage for filming Movies or videos.

Q: Do you think like, Sora could be combined with Apple Vision Pro?

A: Yeah, I can definitely see the use of that. Again, I think both of these two products they have very similar in terms of like just even as a first version the quality are amazing.

Q: Is there any government funding for the development of AI in China?

A: I believe that's the case, yeah. All the major investors are, you could say, like public companies, but not like government or national affiliated, yeah.

Q: So I don't know like how much time does he have to run his company.

A: I feel this is like, it's a very common thing for Chinese professors to also have some startups. So I don't know like how much time you can spend on research versus running the company.

前往小宇宙评论区与主播互动

This podcast episode explores the recent quarterly earnings report from Nvidia, where the company surpassed expectations and broke records. Tracy shares her personal investment strategy, focusing on companies whose products she uses. The hosts discuss the use of GPUs in research and a social experiment that analyzed meeting datasets to determine the types of meetings that lead to multitasking and less attention from attendees. Researchers analyzed a vast dataset of internal Microsoft meetings to identify factors contributing to multitasking, finding that larger meeting sizes, lack of personal relevance, recurring meetings, and morning schedules led to increased multitasking. The study advises scheduling essential meetings in the evening and minimizing group sizes to reduce multitasking and improve meeting effectiveness.

Takeaways

• Nvidia continues its trend of record-breaking quarterly earnings.
• Investing in companies whose products you use can be a solid strategy.
• GPUs are essential for research and development.
• Social experiments can provide valuable insights into human behaviors.
• Larger meeting sizes and recurring meetings tend to lead to increased multitasking.
• The time of day and nature of the meeting can influence multitasking behaviors.
• Important meetings should be scheduled in the evening with smaller group sizes to reduce multitasking.

Outlines

Nvidia's Record-breaking Earnings
00:00NVIDIA's Record-Breaking Earnings: Insights from Tracy and Ziyang
01:03Uncover the Secrets of Multitasking in Meetings: A Data-Driven Analysis
03:06Multitasking in Meetings: Impact of Time and Recurrence
05:38

Q & A

Q: Do you have any like GPUs at home?

A: I don't own any Nvidia stocks because I only buy company stocks when I use the products. I think I definitely use GPU indirectly since I use China GPT. However, I have no more comments since I cannot touch these products.

Q: Can you have a guess like what type of meetings like people will pay less attention to?

A: So based on your own experience i would say maybe when you are not presenting or if everyone is there maybe on like the same job position level you don't have your bosses there then that's just my random guess.

Q: And they also study whether recurring meetings are more distracting or one-off meetings. So based on your own experience, which one will you be more likely to multitask?

A: Probably the current meeting. Because you know the people there well so they probably Careless. They're probably already your friends or people you know well so you just feel more comfortable just doing multitasking stuff there I think.

前往小宇宙评论区与主播互动

前往小宇宙评论区与主播互动

大家好，这是一档《创意玩具》旗下的新节目《科技周会Standing Meeting》。

我会邀请一起创业做AI项目的伙伴们和他/她们的朋友们一起来录制。

节目内容主要是周会闲聊AI和各种业界新闻的部分录音。

主播：Tracy @tracyyxchen , Roger @rogerzhufilm

00:00 介绍Intro

02:15 苹果Vision Pro

07:45 CHI大会历年的赞助商

15:30 参观完DC车展的一些感想

This podcast episode explores the accessibility of VR headsets for glasses wearers, including the introduction of Apple Vision Pro, a collaboration between Apple and a German optics manufacturing company, which offers optical inserts that allow users to wear their prescription lenses inside the VR headset, eliminating the need for glasses. Additionally, the episode discusses the upcoming Apple Vision Pro demo at Apple stores, a prestigious CHI conference, and the state of AI and EV development in China and the rest of the world.

Takeaways

• Apple Vision Pro offers optical inserts that allow glasses wearers to use VR headsets without discomfort.
• The CHI conference received over 4,000 papers with an acceptance rate of approximately 26%, highlighting its high standards.
• China is lagging in AI development but ahead in EV design and development.
• Hyundai IONIQ 5 is considered better than the current Tesla Model 3 and Model Y.
• There is a growing trend of AI-enabled vehicles and fierce competition in the global EV market.

Outlines

Exploring the Accessibility of VR Headsets for Glasses Wearers
00:00Apple Vision Pro: Accessibility Features for Different Vision Conditions
03:58Apple Vision Pro Demo Coming Soon, CHI Conference Acceptance
06:20Exploring the CHI Conference: A Journey Through History and Sponsorships
08:30Evolution of Sponsorships at the Conference: A Data-Driven Analysis
11:02AI and EV: A Global Race
13:54The Rise of AI-Enabled Vehicles and the Global EV Market
17:49The Future of Electric Vehicles: Comparing Features and Designs
21:26

Q & A

Q: What is Apple Vision Pro?

A: Apple Vision Pro is a collaboration between Apple and a German optics manufacturing company that allows people who wear glasses to use VR headsets without having to wear their glasses inside the headset.

Q: What are the different types of eyeglasses you can order for Apple Vision Pro?

A: There are two types of eyeglasses you can order: one for prescription and the other for readers.

Q: What companies are sponsoring the CHI conference?

A: So for this year they have four sponsors. The champion sponsor is Apple and other companies also contributing Adobe Autodesk and Microsoft.

Q: Do you remember what happened to Motorola?

A: I think it was sold to Google once. Yeah, they were sold to Google and now I believe they are a subsidiary branch of Lenovo.

Q: Will Samsung make a car?

A: I haven't heard any news about Samsung making a car, but Sony has a joint venture with Honda to release a car in late 2025 or early 2026.

Q: Which car do you plan to buy after the auto shift?

A: For me, actually, I did take a look at most of the interiors of the cars. I know the trend right now in the Chinese market is they are trying to have bigger and bigger screens in the front. I think Tesla probably started that trend, but all the Chinese manufacturers are trying to one-up them.

前往小宇宙评论区与主播互动