Deprecated: Using null as an array offset is deprecated, use an empty string instead in /home/u876752588/domains/capria.vc/public_html/wp-content/plugins/jet-engine/includes/components/blocks-views/dynamic-content/manager.php on line 113

My usual quarterly updates are for sharing learning from travels and investing across the Global South. In my last mail, I discussed our “Partners’ Views from the [Global South] Ground.” Today, I have some comments on the continued controversy on where GenAI is today and where it’s headed, led by AI curmudgeon Gary Marcus, who I respect but sometimes disagree with.

Breaking with my usual style, I’m starting by sharing some news about our investing activities that I’m particularly proud of. We recently announced an all-cash exit from our portfolio company, Awign, in India. We sold our stake to a leading Japanese Human Resources conglomerate and are beyond pleased to be returning USD 22M to our investors.

I share this exit news as a perfect example of an emerging market VC investment thesis – to find high-quality entrepreneurs early in their startup journey of building scalable solutions for the rising middle class. One area we identified before most investors was the need for sophisticated tech to support the massive demand for qualified gig workers with flexible engagement models, moving them from informal to semi-formal and formal employment. We were among the first to be speaking about “JobTech” in India and have the largest startup portfolio in this sector. That experience has helped us make similar investments in Jobtech in other regions of the Global South. We’ve also been deeply focused on using GenAI in the Jobtech and other sectors to “make already good businesses better” and help our portfolio companies realize that value as they scale their business and impact, as Awign has done exceptionally well.

Please take a minute and read my point/counter-point with Gary Marcus below. And tell me what you think.

Why I am bullish on GenAI and not afraid of the bubble

I follow Gary Marcus and was struck by his recent article in which he discusses his views on “When the GenAI bubble will burst.” His voice certainly helps cut through some of the euphoric froth that we all see in this market. While I do agree we are in a bubble, I do not believe it will burst the way he suggests, and it’s not going to be a 2000 dot-com-like bubble burst or the 2008 financial crisis. While a correction is likely (and needed), I think there is too much significant underlying intrinsic innovation happening today for any kind of bubble burst of that magnitude.

He raises concerns about the ballooning valuation of AI startups and their ability to generate revenue. In some cases, I’m sure he will be proven right. However, there are tangible examples of successful AI implementations driving top-line growth. Generative AI is now contributing revenue to Amazon’s cloud business at an annualized rate equivalent to multiple billions of dollars. Amazon Web Services (AWS) achieved a remarkable milestone by reaching a $100 billion annual revenue run rate. In the first three months of this year, their revenue grew by 17%, which is the fastest it has grown since 2022. Amazon’s customers would not be paying them so much for this infra if they were not seeing a path to generate revenue with it as well.

Gary then discusses the security risks of the technology, particularly perceived threats to cybersecurity with phishing attacks, deep fakes, complex fraud, and automated attacks on the rise. Yes, GenAI will be used by bad guys, but it will be used by the good guys too. Cybersecurity professionals can utilize GenAI tools such as ChatGPT and other large language model (LLM) tools to fortify their system defenses against malicious intrusions. They help businesses analyze massive volumes of log files and network traffic data during cyber incidents, speeding up and automating the response process. Companies that heavily relied on AI and automation saved nearly $1.8 million in breach costs and shortened the time to identify and contain breaches by more than 100 days on average.

Next, Gary brings up AI hallucinations (AI-generated inaccuracies) and the associated challenges in bringing LLMs up to human expert levels for a start. We’ve all experienced hallucinations and inaccuracies; Google users are suffering from this as we speak. OpenAI is considering a new approach to enhance AI’s ability to reason called “process supervision.” Unlike conventional methods that only reward correct final answers, this approach incentivizes AI models for each correct step of reasoning they take. By mimicking a human-like chain of thought, this strategy aims to make AI more explainable and capable of avoiding logical mistakes and fabrications. There are undoubtedly many hundreds, if not many thousands, of highly capable engineers working to address this precise issue across the AI industry. I expect it will get solved much more quickly than it took Microsoft to bring Windows from an experiment to mainstream success.

Gary speaks about how the surge of money flooding into artificial intelligence has resulted in a crypto-like hype that is obscuring incredible scientific progress in the field. Sure, there is hype, but plenty is for good reason. In a newly shared video that everyone should watch, Reid Hoffman (LinkedIn’s founder) presents REID AI – his AI counterpart. This AI, supported by a specialized chatbot trained on Reid’s vast content and an amazing video clone of Reid, engages in a Q&A session with him. During this exchange, REID AI challenges Reid’s perspectives on AI and technology, offering fresh insights into AI’s ability to mimic human interaction and foster critical thinking. You will be impressed by the implications of video chatbot technology, which promises positive applications across various industries. Yes, with drawbacks too, but I think the positives outweigh the negatives.

My take:

Stay calm and invest with due care in applied Generative AI, and invest with caution in AI infrastructure (especially via the incumbents). There is plenty of value being created and money to be made.