Logo
    Search

    #137 - Salesforce Copilot, Chip Crunch, Meta Rival to ChatGPT, AI for paralysis patients

    enSeptember 20, 2023

    Podcast Summary

    • Skepticism about Salesforce's new AI assistant, Einstein Co-pilotDespite Salesforce's new AI feature, Einstein Co-pilot, across Slack and Tableau, doubts remain about its usefulness for complex tasks and edge cases.

      Salesforce has recently launched its AI assistant, Einstein Co-pilot, across its apps, including Slack and Tableau. This new feature aims to provide personalized answers to customer questions, generate emails, marketing campaigns, and summarize video calls. However, during our discussion, we expressed skepticism about the usefulness of these AI features, particularly in the context of Slack. We shared our experiences with other AI-powered tools and bots, expressing doubts about their ability to handle complex tasks or edge cases. Despite the growing trend of companies integrating AI co-pilots into their software, we remain uncertain about their effectiveness and usefulness, especially for more complex issues. While they may be helpful for basic instructions or triaging, we believe that human expertise and interaction will continue to be essential for handling complex tasks in various industries.

    • AI is transforming industries and platformsAI is being integrated into various industries and platforms, offering insights, automation, and improved user experiences, with Salesforce, Roblox, and Bybit being recent examples

      Artificial intelligence (AI) is increasingly being integrated into various industries and platforms, making processes more efficient and accessible. Salesforce, a leading CRM company, has been investing in AI for quite some time and is likely implementing it effectively. Roblox, a popular gaming platform, has recently introduced an AI chatbot, Roblox Assistant, to help users build virtual worlds. This assistant can modify game scenes based on user requests, making it easier for users to create custom versions of games without requiring extensive technical skills. Bybit, a crypto currency exchange, has debuted an AI trading assistant called trade GPT, which provides market analysis and answers questions using AI. These developments suggest that AI is becoming an essential tool for businesses and individuals in various sectors, offering insights, automation, and improved user experiences. While there may be concerns about the accuracy and impact of AI-driven recommendations, the overall trend indicates a future where AI plays a significant role in shaping industries and user experiences.

    • Significant advancements in AI technology by Chad GPT, Ant Group, and TSMCChad GPT offers trading insights, Ant Group's models conduct investment analysis, and TSMC faces chip crunch affecting AI chip production. AI adoption raises questions about financial market ecosystem correlations and affordability due to resource competition.

      There are significant advancements in AI technology being made by various companies, including Chad GPT, Ant Group, and TSMC. Chad GPT is offering trading insights and advice, while Ant Group's CCZU 1.0 and 2.0 models can conduct investment analysis and reasoning for finance professionals and consumers, respectively. TSMC has warned of an ongoing chip crunch that will last another 18 months due to advanced packaging capacity constraints, affecting the production of high-end GPUs and other AI-oriented chips. The widespread adoption of AI technology raises questions about potential correlations and failures within the financial market ecosystem. Despite these advancements, there are challenges, such as regulatory approval and chip scarcity, that may impact the accessibility and affordability of these technologies for some organizations and individuals. The recent influx of money into AI is leading to increased competition for limited resources, making it a significant challenge for the industry.

    • Bing's Impact on Google's Search Dominance MinimalDespite Bing's release, Google's search dominance remains due to users preferring to click and verify results, and chatbots' impact on search being minimal.

      While Microsoft's Bing chatbot release was highly anticipated to challenge Google's search dominance, the impact has been minimal. Google, being late to the chatbot race, didn't lose significant market share, with Bing only gaining a 1% increase to a 3% worldwide market share. Google's search quality, which has been declining according to some, may have provided an opportunity for an alternative. However, chatbots like Bing's and OpenAI's ChatGPT may not be the best interface for search as users often still prefer to click and verify results themselves. Instead, it seems that chatbots are impacting other sources of information, such as Wikipedia, as they can answer random questions effectively. OpenAI, which makes money through its API and having developers build on top of it, is hosting its first developer conference on November 6, further emphasizing its commitment to this product. While I personally haven't found joy or use in chatbots for my needs, they may excel in simpler, hand-built solutions and internal tooling. The future of search and chatbots remains to be seen, but for now, it appears that Google's search dominance remains intact.

    • AI's role in education and financeOpenAI hosting hackathon, Elo raises funds for AI reading coach, NASDAQ launches first AI-driven order type, AI making a difference in industries, but complexities and limitations remain

      There are ongoing developments in the integration of artificial intelligence (AI) in various industries, from education to finance. At an upcoming event, OpenAI may be hosting a hackathon where AI could play a significant role. In the education sector, a startup called Elo has raised $15 million for its AI reading coach service, which delivers personalized tutoring to children using AI and speech recognition technology. In finance, NASDAQ has received SEC approval to launch the first AI-driven order type, which uses reinforcement learning to make real-time adjustments in the trading process. These developments demonstrate the potential of AI to make a difference in various industries, even if it may not replace human intervention entirely. However, it's important to remember that while these advancements are promising, they also come with complexities and limitations that require further exploration. For instance, it's unclear how Elo measures the progress of its AI tutoring system, and the specifics of NASDAQ's AI-driven order type remain to be seen. Nonetheless, these developments represent an exciting step forward in the application of AI to solve real-world problems.

    • NVIDIA's Tensor RT dash LLM: Doubling the Speed of Large Language ModelsNVIDIA's Tensor RT dash LLM offers twice the speed of large language models without requiring hardware upgrades, thanks to optimized batching processes for AI workloads.

      NVIDIA's Tensor RT dash LLM is significantly improving the performance of large language type models like Llama and Falcon, offering twice the speed without requiring hardware upgrades. This is a result of optimized batching processes for AI workloads, which streamline scheduling and reduce bottlenecks. NVIDIA's dominance in this field is partly due to their long-standing CUDA software and recent addition of TensorRT. Meta, on the other hand, is reportedly developing an open-source generative AI system rumored to be more powerful than GPD4, aiming to expand access to advanced models. However, the debate continues on whether these open-source initiatives truly democratize AI technology or just centralize it further, as they still require significant resources for implementation.

    • Open sourcing of AI models and researchCompanies and individual developers are open-sourcing AI models and research, with Indian developers leading the way in fine-tuning high-ranking models. Concerns about benchmark effectiveness and potential leakage exist, but companies like Accubit Technology and venture capital firms like A16Z continue to support open source development.

      The open sourcing of AI models and research is gaining momentum, with companies like Meta and individual developers contributing to this trend. The Indian developers' success in fine-tuning models like Gen Z70B, which ranks high on Hugging Face leaderboards, showcases this proliferation. However, there are concerns about the effectiveness of benchmarks for evaluating large language models and the potential for benchmark leakage. Companies like Accubit Technology are also open-sourcing their language models, showcasing their capabilities to potential clients. Venture capital firms, such as Andreessen Horowitz (A16Z), are further encouraging open source development by offering grants to developers. This trend signifies a shift towards collaborative AI research and development.

    • Restoring Communication for Paralyzed Patients with AIAI technology is revolutionizing communication for paralyzed individuals by translating brain signals into words and voice.

      The latest advancements in Artificial Intelligence (AI) are making a significant impact on various fields, particularly in helping people with disabilities communicate. A prime example is the research conducted at the University of California, which has resulted in a device that restores communication for paralyzed patients by translating their brain signals into words and voice through AI technology. This is a transformative development, as it enables individuals with limitations to express themselves and engage with the world. However, there are ethical considerations to be addressed, such as the use of avatars for communication, which may raise concerns. The researchers have reported rapid decoding speeds and reasonable accuracy, but it's unclear how novel their findings are due to the paper being published in a prestigious research venue, Nature, which requires a subscription to access. Overall, this innovation represents a crucial step forward in the realm of assistive technology and accessibility.

    • Exploring ways to improve language models and healthcare technologyResearchers develop efficient benchmarking methods, optimize datasets, and create AI-based tools for remote Parkinson's disease assessments, enhancing efficiency, accessibility, and reliability.

      Researchers are constantly exploring ways to improve the efficiency and reliability of language models through innovative benchmarking methods and dataset optimization. For instance, a recent paper focuses on efficient benchmarking of language models, allowing for quicker evaluations with less data usage. Another project, Helm, aims to reduce computation costs by up to 200 times while maintaining reliability and correct model rankings. Furthermore, researchers at the University of Rochester have developed an online AI-based test for Parkinson's disease severity, enabling remote assessments within minutes using a webcam and tap test, ultimately making the process more accessible and convenient for patients. Overall, these advancements in language model evaluation and healthcare technology demonstrate the ongoing commitment to improving efficiency, accessibility, and reliability in various fields.

    • AI's role in healthcare and quantum computingAI shows promise in healthcare for complementing human experts in evaluating symptoms, while in quantum computing, reinforcement learning optimizes quantum error correction.

      Artificial intelligence (AI) is showing promise in various fields, including healthcare and quantum computing. In the healthcare sector, a study compared Virtual Assistant Intelligence (VAI) systems to human experts in evaluating Parkinson's disease symptoms. While human experts outperformed AI models, they did not always agree among themselves, highlighting the potential value of AI as a complementary tool. In quantum computing, scientists have used machine learning, specifically reinforcement learning, to optimize quantum error correction, which is crucial for maintaining the consistency of quantum states. The use of RL in this context might be due to the trial-and-error nature of optimization when a full system model cannot be obtained. Overall, these developments underscore the growing role of AI in advancing scientific research and potential applications. However, it's important to remember that these are preliminary findings and further research is necessary to validate and refine these approaches.

    • AI Regulation Discussed in US Senate HearingThe US Senate held a six-hour hearing on AI, discussing potential uses, concerns, and the need for regulation. Senator Schumer proposed a new regulator, while the first legal ruling on AI-generated artwork was made.

      Last week's six-hour AI hearing in the US Senate, attended by tech leaders like Musk, Zuckerberg, and Gates, as well as other industry experts and civil society representatives, highlighted the need for government regulation in AI. Discussions ranged from the potential uses of AI to solve world hunger to the concern over powerful open-source models. Senator Schumer proposed the idea of a new regulator for AI, and while the session might have been seen as a Q&A for lawmakers, it also allowed industry players to advocate for their preferred future of AI. However, the closed session prevented reporters from sharing specific details. Meanwhile, a US district judge made history by ruling that AI-generated artwork cannot be copyrighted, marking the first legal finding on the subject. The implications of these developments for AI policy and public perception are yet to be seen.

    • AI-generated art and copyright lawA recent court ruling may not allow copyright for AI-generated art, but guidelines and proposed regulations may impact text image models and human-algorithm contributions.

      A recent court ruling may have set a legal precedent for not being able to copyright AI-generated art, as seen in the case of Stephen Thaler and his AI-generated artwork. However, this might not apply to text image models. The US Copyright Office has also published guidelines for AI artwork applicants to indicate human and algorithm contributions. Two US senators have proposed a regulatory framework for AI, which includes licensing, auditing, and liability for privacy and civil rights violations. The US Copyright Office has launched a study on AI and copyright law, inviting public comment to inform Congress and guide regulatory endeavors. These developments suggest that there will be continued scrutiny and regulation of AI in the legal and policy spheres.

    • AI's Impact on Film: Enhancing Effects or Infringing Rights?AI is transforming film production by creating visual effects, de-aging actors, and even generating movies, but its use raises concerns about copyrightability, transparency, and infringement.

      The intersection of artificial intelligence (AI) and the creative industry, particularly film, is becoming more prominent. This was evident at the Venice Film Festival where AI was used to create visual effects, de-age actors, and even generate entire films. However, the use of AI in this way raises questions about copyrightability, transparency, and infringement. The public feedback on these issues is unclear in its impact, but it is important that the conversation is happening. The use of AI in film is currently being explored as a tool for enhancing special effects and filtering, rather than replacing human labor or creating entire movies. It remains to be seen how this technology will evolve and what its implications will be for the creative industry as a whole.

    • AI as a tool for smaller studiosAI is helping smaller studios enhance creativity, streamline processes, and improve efficiency without replacing human workers.

      AI technology is increasingly being adopted by small studios and developers to streamline their processes and enhance their creative output, without necessarily replacing human workers. Revolution Software, for instance, is using AI to upscale the animation in their remake of the old game Broken Sword, allowing their animators to focus on more complex tasks. Similarly, the developer of the Terraforming Mars board game used AI to generate some art assets for their new Kickstarter project, enabling their team to produce more assets more quickly. These examples show that AI can be a useful tool for smaller studios to improve efficiency and productivity, rather than a threat to human jobs. However, it's important for studios to be transparent about their use of AI and address any concerns or controversies that may arise.

    Recent Episodes from Last Week in AI

    #173 - Gemini Pro, Llama 400B, Gen-3 Alpha, Moshi, Supreme Court

    #173 - Gemini Pro, Llama 400B, Gen-3 Alpha, Moshi, Supreme Court

    Our 173rd episode with a summary and discussion of last week's big AI news!

    With hosts Andrey Kurenkov (https://twitter.com/andrey_kurenkov) and Jeremie Harris (https://twitter.com/jeremiecharris)

    See full episode notes here.

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form.

    Email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

    In this episode of Last Week in AI, we explore the latest advancements and debates in the AI field, including Google's release of Gemini 1.5, Meta's upcoming LLaMA 3, and Runway's Gen 3 Alpha video model. We discuss emerging AI features, legal disputes over data usage, and China's competition in AI. The conversation spans innovative research developments, cost considerations of AI architectures, and policy changes like the U.S. Supreme Court striking down Chevron deference. We also cover U.S. export controls on AI chips to China, workforce development in the semiconductor industry, and Bridgewater's new AI-driven financial fund, evaluating the broader financial and regulatory impacts of AI technologies.  

    Timestamps + links:

    Last Week in AI
    enJuly 07, 2024

    #172 - Claude and Gemini updates, Gemma 2, GPT-4 Critic

    #172 - Claude and Gemini updates, Gemma 2, GPT-4 Critic

    Our 172nd episode with a summary and discussion of last week's big AI news!

    With hosts Andrey Kurenkov (https://twitter.com/andrey_kurenkov) and Jeremie Harris (https://twitter.com/jeremiecharris)

    Feel free to leave us feedback here.

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekinai.com and/or hello@gladstone.ai

    Last Week in AI
    enJuly 01, 2024

    #171 - - Apple Intelligence, Dream Machine, SSI Inc

    #171 - - Apple Intelligence, Dream Machine, SSI Inc

    Our 171st episode with a summary and discussion of last week's big AI news!

    With hosts Andrey Kurenkov (https://twitter.com/andrey_kurenkov) and Jeremie Harris (https://twitter.com/jeremiecharris)

    Feel free to leave us feedback here.

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + Links:

    Last Week in AI
    enJune 24, 2024

    #170 - new Sora rival, OpenAI robotics, understanding GPT4, AGI by 2027?

    #170 - new Sora rival, OpenAI robotics, understanding GPT4, AGI by 2027?

    Our 170th episode with a summary and discussion of last week's big AI news!

    With hosts Andrey Kurenkov (https://twitter.com/andrey_kurenkov) and Jeremie Harris (https://twitter.com/jeremiecharris)

    Feel free to leave us feedback here.

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + Links:

    Last Week in AI
    enJune 09, 2024

    #169 - Google's Search Errors, OpenAI news & DRAMA, new leaderboards

    #169 - Google's Search Errors, OpenAI news & DRAMA, new leaderboards

    Our 168th episode with a summary and discussion of last week's big AI news!

    Feel free to leave us feedback here: https://forms.gle/ngXvXZpNJxaAprDv6

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + Links:

    Last Week in AI
    enJune 03, 2024

    #168 - OpenAI vs Scar Jo + safety researchers, MS AI updates, cool Anthropic research

    #168 - OpenAI vs Scar Jo + safety researchers, MS AI updates, cool Anthropic research

    Our 168th episode with a summary and discussion of last week's big AI news!

    With guest host Gavin Purcell from AI for Humans podcast!

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + Links:

    Last Week in AI
    enMay 28, 2024

    #167 - GPT-4o, Project Astra, Veo, OpenAI Departures, Interview with Andrey

    #167 - GPT-4o, Project Astra, Veo, OpenAI Departures, Interview with Andrey

    Our 167th episode with a summary and discussion of last week's big AI news!

    With guest host Daliana Liu (https://www.linkedin.com/in/dalianaliu/) from The Data Scientist Show!

    And a special one-time interview with Andrey in the latter part of the podcast.

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + links:

    Last Week in AI
    enMay 19, 2024

    #166 - new AI song generator, Microsoft's GPT4 efforts, AlphaFold3, xLSTM, OpenAI Model Spec

    #166 - new AI song generator, Microsoft's GPT4 efforts, AlphaFold3, xLSTM, OpenAI Model Spec

    Our 166th episode with a summary and discussion of last week's big AI news!

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + links:

    Last Week in AI
    enMay 12, 2024

    #165 - Sora challenger, Astribot's S1, Med-Gemini, Refusal in LLMs

    #165 - Sora challenger, Astribot's S1, Med-Gemini, Refusal in LLMs

    Our 165th episode with a summary and discussion of last week's big AI news!

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + links:

    Last Week in AI
    enMay 05, 2024

    #164 - Meta AI, Phi-3, OpenELM, Bollywood Deepfakes

    #164 - Meta AI, Phi-3, OpenELM, Bollywood Deepfakes

    Our 164th episode with a summary and discussion of last week's big AI news!

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

    Timestamps + links:

    Last Week in AI
    enApril 30, 2024

    Related Episodes

    EP 130: Build a Brand Live with AI (Part 1)

    EP 130: Build a Brand Live with AI (Part 1)

    Can you build an entire brand with AI? What capabilities do ChatGPT and its plugins offer to help go from nothing to a full brand? We're building a brand from scratch using the live audience for help and suggestions. From finding our brand name and identity to brand taglines and logos, we're kicking off part 1 of this brand building with a bang. 

    Newsletter: Sign up for our free daily newsletter
    More on this Episode: Episode Page
    Join the discussion: Ask Jordan questions about AI and brand building
    Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
    Website: YourEverydayAI.com
    Email The Show: info@youreverydayai.com
    Connect with Jordan on LinkedIn

    Timestamps:
    [00:02:00] Daily AI news
    [00:06:00] Creating a real world product idea
    [00:11:15] Brand names using ChatGPT and plugins
    [00:24:00] Competitive analysis and target audience for your brand
    [00:34:00] Developing marketing language with ChatGPT
    [00:44:45] Designing a logo with DALL-E 3 inside ChatGPT
    [00:59:00] Closing thoughts for part 1

    Topics Covered in This Episode:
    1.  Leveraging AI for Brand Building
    2. Using ChatGPT for marketing research and copy
    3. Designing logos with DALL-E 3 inside ChatGPT

    Keywords:
    brand building, AI, ChatGPT, internet browsing, plug-ins, task automation, research, reading articles, analyzing competitors, AI news, newsletter, Google Q3 earnings call, ad business, trends, up-to-date information, mid journey, DALL-E, logo design, minimalistic, futuristic, digital nomads, live demonstration, natural language processing, podcast episode, audience participation, artificial intelligence, search generative experience, AI-powered ads, chatbots, consumer product good, ecommerce product, marketing language, product creation

    The 10 Biggest AI Events of 2023

    The 10 Biggest AI Events of 2023
    NLW builds off of this list of the 10 biggest events of AI of 2023 to do a yearly recap! Interested in the January AI Education Beta program? Learn more and sign up here - https://bit.ly/aibeta ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

    Anthropic Raising at a $20B-$30B Valuation

    Anthropic Raising at a $20B-$30B Valuation
    Just a couple weeks after announcing their big Amazon deal, Anthropic is out fundraising again with a $2B round that is reported to have Google's involvement and is seeking a $20B-$30B valuation. Before that on the Brief, a Mr. Beast deepfake ad scam; Morgan Stanley says 40% of workers impacted by AI in the next 3 years, and more. TAKE OUR SURVEY ON EDUCATIONAL AND LEARNING RESOURCE CONTENT: https://bit.ly/aibreakdownsurvey ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

    #140 - Yasa vs ChatGPT, Waymo expands, scaling robot learning, AI watermarks

    #140 - Yasa vs ChatGPT, Waymo expands, scaling robot learning, AI watermarks

    Our 140th episode with a summary and discussion of last week's big AI news!

    Read out our text newsletter and comment on the podcast at https://lastweekin.ai/

    Email us your questions and feedback at contact@lastweekin.ai

    Note: the CEO of GitHub disputes "Report: GitHub Copilot Loses an Average of $20 Per User Per Month"

    Timestamps + Links: