Kabir's Tech Dives
Download on the App Store

Kabir's Tech Dives episodes

  • Should AI Run Your Computer? Anthropic Thinks So!

    Anthropic has announced two new AI models: Claude 3.5 Sonnet, which has been significantly upgraded, especially in coding abilities, and Claude 3.5 Haiku, a faster model that surpasses even the company's previous largest model in many areas. In a groundbreaking move, Anthropic has also introduced computer use, a new feature allowing developers to direct Claude to use computers as people do, enabling the AI to interact with computer interfaces, navigate websites, and complete complex tasks. While still in its early stages, computer use has the potential to revolutionize how people work with AI, automating processes and making it possible for AI to perform tasks that were previously impossible.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    13 min
  • Climate Tech Startup Opportunities

    This episode is based on an annual report from Breakthrough Energy and examines the state of the clean energy transition and Breakthrough Energy's efforts to accelerate it. 

    The report outlines the organization's climate tech discovery, development, and deployment initiatives. It highlights the increasing interest from investors and corporations in climate tech and the crucial need for policy changes, especially regarding transmission infrastructure, to enable the transition to a clean energy economy. 

    The report also showcases several Breakthrough-backed companies developing solutions for decarbonizing various sectors such as agriculture, manufacturing, transportation, and buildings.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    35 min
  • Colossus Super Computer is ON | Elon Musk

    This episode discusses an article from HPCwire about Elon Musk's xAI project, which has created a massive AI training cluster called Colossus. The article details the cluster's size, power requirements, and potential impact on AI.

    It also includes information about Nvidia's contribution to the project through its H100 and H200 GPUs, the use of mobile generators to provide power, and the estimated project cost.

    Original Article

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    7 min
  • Can Mechanical Battery Outperform Li-ION | QNETIC

    This episode is about the mechanical battery startup Qnetic, a company that has developed a new flywheel energy storage (FESS) system called Qnetic.

    This system uses a lightweight, high-strength fiber rotor and novel levitation bearings to achieve a large storage capacity and long discharge duration, making it ideal for storing renewable energy. The Qnetic system is designed to be safe, cost-effective, and scalable for grid-level applications, enabling the transition to a clean energy future.

    Qnetic Website
    https://www.qnetic.energy

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    11 min
  • How to Grow Your Non-Tech Startup with AI

    This episode is about HubSpot and provides a comprehensive overview of large language models (LLMs) and how business owners can leverage their power. The guide outlines four levels of LLM usage: prompt engineering, prompt engineering with APIs, fine-tuning, and building your LLM. 

    It emphasizes the importance of input and output, highlighting how well-crafted prompts lead to better results. The guide also explores techniques like few-shot, chain-of-thought, and self-consistency prompting to improve LLM performance. Finally, it discusses the cost and benefits of each level and emphasizes that even small businesses can benefit from using LLMs to automate tasks, analyze data, and generate content.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    15 min
  • How to Win with AI Contents as a Startujp

    This episode focuses on a report from MasterDai, a content marketing agency, examines the use of artificial intelligence (AI) in content marketing and search engine optimization (SEO). The report analyzes a survey of over 2,600 small and medium businesses to determine the benefits and challenges of using AI for content creation. The authors explore the results of using AI for various tasks, including generating content, rewriting text, and optimizing content for SEO. The report also presents insights from leading content marketing experts on the impact of AI on the industry. The report concludes that AI can be a valuable tool for content marketing, but human oversight is crucial for quality and effective results.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    12 min
  • AI Bug Bounty is Here | 0Din Portal

    The 0Din Portal is a new platform designed to simplify the process of finding and reporting security vulnerabilities in Generative AI (GenAI) systems. This platform aims to make bug bounty hunting more efficient and rewarding for researchers, and provides tools for submitting and managing vulnerabilities, as well as a variety of learning resources such as blogs, tutorials, and Capture-the-Flag (CTF) competitions. The platform is built with the goal of fostering a collaborative community focused on making AI safer and more secure.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    22 min
  • DEEP FAKERY—An Action Plan

    This episode deals with a white paper, the product of a workshop with experts from various fields, examining the burgeoning threat of deep fakery.

    This technology allows for the creation of compelling yet fabricated audio and video content. The paper outlines the multiple harms deep fakery poses, including reputational damage, emotional distress, and the erosion of trust in media. It then analyzes incentives that drive the creation and spread of deep fakes, including financial gain, social influence, and political manipulation.

    The authors propose a plan of action encompassing technical, socio-economic, and legal solutions, emphasizing the need for collaborative research to combat this evolving threat. The paper calls for developing detection algorithms, authentication protocols, and robust legal frameworks to address the dangers of deep fakery.

    Original Whitepaper

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    12 min
  • AI Deepfake Detection Using FakeCatcher by Intel

    Intel is developing a real-time deep fake detection platform called FakeCatcher. It uses photoplethysmography (PPG), which measures blood flow changes in the face, and eye-gaze detection to identify fake videos with 96% accuracy. This technology is crucial to combating harmful deep fakes used for political misinformation, impersonation, and adult content. 

    This technology signifies a new era of understanding what makes us human and authenticating our digital identities.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    4 min
  • How & Why Use ChatGPT Canvas | OpenAI

    Canvas is a new feature for ChatGPT that allows users to work on writing and coding projects more interactively. It provides a dedicated workspace where users can edit text or code, get suggestions from ChatGPT, and collaborate on projects. Canvas is designed to improve the context of interactions, providing ChatGPT with a better understanding of the user's intent. 

    Users can access Canvas through the ChatGPT web interface by selecting GPT-4o with Canvas from the model selector. Canvas offers various shortcuts for editing text and code, including suggestions for improvements, adjusting text length, changing the reading level, adding emojis, and reviewing code. 

    Users can also highlight specific sections of text or code to ask ChatGPT for edits or questions about those sections. While Canvas currently does not support image uploads, files, GPTs, or shared conversations, these features are expected to be available soon.

    Send us a text


    Podcast:
    https://kabir.buzzsprout.com


    YouTube:
    https://www.youtube.com/@kabirtechdives

    Please subscribe and share.

    7 min

About Kabir's Tech Dives

From the publisher's feed

I'm always fascinated by new technology, especially AI. One of my biggest regrets is not taking AI electives during my undergraduate years. Now, with consumer-grade AI everywhere, I’m constantly…