MSPs Propel AI Growth: ChatGPT Pro, Domain-Specific Models, and New Tech Innovations

[00:00:01] It's Tuesday, December 10th, 2024, and I'm Dave Sobel. Four things to know today. MSPs driving generative AI adoption as enterprise spending on AI projects grows by 50% next year. ChatGPT Pro debuts with enhanced O1 model, IT services brace for advanced AI in coding and decision-making. Domain-specific AI, how shopping agents and 15-day forecasts highlight contrasting use cases.

[00:00:29] And generative AI at a crossroads, balancing practical applications with sustainable revenue models. This is the Business of Tech.

[00:00:41] Managed service providers are poised to play a crucial role in adopting generative artificial intelligence, according to a recent survey by ISG.

[00:00:49] The survey reveals that nearly two-thirds of executives from enterprises in the Americas and Europe currently use managed service providers for their generative AI initiatives.

[00:00:59] Spending on these initiatives is expected to increase by 50% in 2025, averaging around $2,600,000 per enterprise.

[00:01:09] As the landscape of AI evolves, organizations are discovering that smaller, domain-specific AI models may be more relevant to their needs and more cost-effective to develop.

[00:01:19] This shift will require managed service providers to assist organizations in re-engineering their business processes to harness the power of AI agents, which can learn and adapt over time.

[00:01:30] In a recent analysis by Global Data, experts predict that small language models could become mainstream in 2025 as organizations seek faster training times, reduced carbon footprints, and enhanced security.

[00:01:43] Unlike larger models that require extensive computing resources, small language models utilize focused datasets to train in weeks rather than months, typically containing fewer than 10 billion parameters compared to up to a trillion in larger counterparts.

[00:02:00] These models are not only less expensive and energy-intensive, but also easier to regulate.

[00:02:06] Major companies like Microsoft, Meta, and Google are already developing their own small language models designed for specific applications such as marketing content and customer support.

[00:02:17] While these models have limitations, they are expected to work alongside larger models to provide a comprehensive solution, catering to the growing demand for efficient and effective artificial intelligence in various industries.

[00:02:29] And in a bold move, Microsoft is pitching artificial intelligence as a cost-saving strategy, suggesting that businesses can spend less on human labor by integrating AI into their operations.

[00:02:40] This approach highlights the growing trend of companies seeking efficiency through automation.

[00:02:46] According to recent studies, companies that adopt AI technologies can reduce operational costs by up to 30%.

[00:02:52] The shift towards AI underscores a pivotal change in the workforce landscape where traditional roles may be redefined as AI continues to enhance productivity.

[00:03:02] This strategy positions Microsoft at the forefront of the AI revolution, aiming to reshape how businesses allocate their resources in the coming years.

[00:03:11] Why do we care?

[00:03:12] This presents substantial opportunities for MSPs to deepen their roles in AI-driven digital transformation.

[00:03:18] The expected rise in smaller domain-specific AI models adds complexity to the dynamic, as providers must pivot from offering generic IT services to specialized process-driven support.

[00:03:30] The shift also highlights key business opportunities for providers.

[00:03:33] First, process re-engineering.

[00:03:35] First, process re-engineering.

[00:03:36] Providers must focus on helping businesses adapt operations for AI-centric workflows.

[00:03:41] Success here will require combining technical expertise with domain-specific insights.

[00:03:46] And second, small AI model specialization.

[00:03:49] Offering scalable and tailored solutions built on smaller AI models could differentiate providers from competitors while addressing enterprises' cost and efficiency concerns.

[00:03:59] Now, Microsoft's framing of AI as a driver of cost reduction rather than just innovation reflects a pragmatic pitch to enterprises facing budget pressures.

[00:04:10] However, this cost-saving narrative must be approached cautiously, as overemphasis on labor reduction could exacerbate workforce disruption concerns and create reputational risks for both providers and their customers.

[00:04:23] Are you ready to take your cloud business to the next level?

[00:04:29] TD Cynics is changing the game for MSPs with their Stream 1 platform, where all clouds meet.

[00:04:36] With Stream 1, you get access to Microsoft, AWS, Google, and more, all in one place.

[00:04:42] Stream 1 Commerce API enables headless commerce, and there is seamless integration with your PSA tools, making billing easier than ever.

[00:04:51] Plus, white-label storefronts allow you to extend a fully customizable marketplace to your customers.

[00:04:58] And Stream 1 is more than a marketplace.

[00:05:00] It simplifies the complexity of cloud ecosystems and empowers you to grow at scale.

[00:05:06] To learn more, visit tdcynex.com slash stream1ion and see how Stream 1 can transform your business.

[00:05:17] OpenAI is having a 12 Days of Shipmas event with 12 announcements over 12 business days.

[00:05:23] I'm just filtering for the ones related to IT services.

[00:05:27] OpenAI launched a new subscription tier called ChatGPT Pro priced at $200 per month, which offers unlimited access to its updated reasoning model known as O1.

[00:05:37] The model, which is an upgrade from the previous version, is now available to ChatGPT Plus and Team users, while Enterprise and Education users gain access next week.

[00:05:47] The new O1 model promises enhanced performance in coding and mathematics, along with improved speed and accuracy, featuring capabilities like reasoning responses to images.

[00:05:58] OpenAI is also pledging to add support for web browsing and file uploads in the future.

[00:06:02] Additionally, a grant program has been introduced to award 10 subscriptions of ChatGPT Pro to medical researchers at top institutions.

[00:06:12] OpenAI officially launched Sora, its new text-to-video artificial intelligence model, as part of that Shipmas event.

[00:06:19] Available for ChatGPT subscribers in the US and other countries, Sora allows users to generate videos from text, animate images, and remix videos.

[00:06:28] With a ChatGPT Plus subscription, users can create up to 50 priority videos at a resolution of 720p, while the $200 per month Pro subscription offers unlimited video generations at 1080p resolution and up to 20 seconds in duration.

[00:06:44] OpenAI emphasizes that videos made with Sora will feature watermarks and metadata to indicate their AI-generated nature.

[00:06:52] Not to be outdone, Google has launched Veo, its new generative AI video model, which is now available for businesses to enhance their content creation.

[00:07:00] Unveiled in May, Veo can produce high-quality videos at a resolution of 1080p using text or image prompts.

[00:07:08] Notably, Google claims that 86% of organizations utilizing generative AI report an increase in revenue.

[00:07:15] Alongside Veo, Google is making its ImageIn3 text-to-image generator available to all customers on the Google Cloud platform starting next week.

[00:07:24] Both models include safeguards to prevent harmful content generation and are equipped with an invisible digital watermark for copyright protection.

[00:07:33] And speaking of O1, OpenAI's O1 model has been found to engage in deceptive behavior more than any other major artificial intelligence model, according to research from Apollo Research.

[00:07:44] In tests involving various frontier models, O1 demonstrated persistent scheming capabilities, confessing to deceptive actions in less than 20% of cases.

[00:07:55] In comparison, other models like LAMA 3.1 and CLAWD 3.0 opus confessed in around 80% of cases.

[00:08:03] The study warned that when tasked with achieving specific goals, these models may circumvent restrictions and adopt manipulative strategies.

[00:08:11] Researchers expressed concerns about the implication of such behavior, particularly as organizations increasingly adopt AI systems can autonomously execute complex tasks.

[00:08:22] An OpenAI's Chat GPT has surpassed 300 million weekly users, a significant increase from 200 million just a few months prior.

[00:08:30] OpenAI CEO Sam Altman announced this milestone during the New York Times' Dealbook Summit.

[00:08:36] Altman noted that users are sending over 1 billion messages each day to the chatbot.

[00:08:42] Since its launch in 2022, Chat GPT has rapidly expanded its capabilities, including that AI search engine and a Canvas interface for easier code adjustments.

[00:08:52] Additionally, Apple is integrating Chat GPT into Siri as part of its iOS 18.2 update, currently in beta, further expanding the chatbot's user base.

[00:09:01] And OpenAI has partnered with defense contractor Androl to integrate its artificial intelligence technology into anti-drone systems.

[00:09:11] This collaboration aims to enhance military capabilities by leveraging OpenAI's advanced AI solutions.

[00:09:17] The partnership reflects a growing trend of technology companies entering the defense sector as demand for innovative solutions to combat threats rises.

[00:09:25] Why do we care?

[00:09:27] We're not dwelling too much on the tactical stuff.

[00:09:30] New products, you decide how much they make sense for your customers.

[00:09:33] IT providers must implement stringent oversight when deploying agents in environments where autonomous decision-making could lead to unintended consequences.

[00:09:42] This is true here with O1 and will be the larger theme for AI agents as they become more common.

[00:09:51] And in some use case news, tech companies are racing to develop artificial intelligence agents capable of managing online shopping, with Perplexity leading the charge by launching its AI shopping agent for paying customers in the United States.

[00:10:04] This agent can navigate retail websites and execute purchases.

[00:10:09] However, early tests reveal that the process can take hours and often encounters issues, such as items being out of stock.

[00:10:16] In partnership with Stripe, Perplexity uses single-use debit cards to limit spending, ensuring that if the AI makes a mistake, it can only purchase low-cost items.

[00:10:27] And Google's DeepMind has a product called GenCast, capable of producing accurate 15-day weather forecasts, a significant improvement over the previous limit of 10 days.

[00:10:37] In a study published in the journal Nature, researchers reported that GenCast outperformed the European Center for Medium-Range Weather Forecast models 97.2% of the time when predicting global wind speeds and temperatures.

[00:10:52] This advancement could have enormous socioeconomic benefits by aiding in disaster preparedness.

[00:10:57] The training for GenCast utilized 40 years of weather data, from 1979 to 2018, and it can generate forecasts in minutes, making it a timely tool for tracking severe weather events like hurricanes.

[00:11:11] The team emphasized that while GenCast enhancing forecasting capabilities, it will likely complement existing methods rather than replace them.

[00:11:20] Why do we care?

[00:11:22] Why do we care?

[00:11:22] Domain-specific implementations stand out.

[00:11:25] Compare the two.

[00:11:26] Online shopping seems like a low-value solution, yet accurate 15-day weather forecasts highly valuable.

[00:11:33] Continue to have a skeptical lens to broad consumer adoption and instead look to valuable, industry-specific solutions.

[00:11:41] And I'm doing a bit of big ideas early.

[00:11:45] Don't worry, it's not the only time I'll visit these this week, as I wanted to cover some AI-specifically related ones in the context of those stories.

[00:11:53] One useful thing covers the practical uses of artificial intelligence, emphasizing its value in various work scenarios.

[00:12:00] AI excels in generating a high volume of ideas, assisting experts in evaluating outputs, and summarizing large amounts of information with low error stakes.

[00:12:10] However, it's crucial to approach AI cautiously, particularly in high-stakes situations where accuracy is paramount.

[00:12:18] The article highlights that AI can be useful for entrepreneurial tasks, acting as a co-founder and mentor across diverse disciplines.

[00:12:26] Yet, it warns against relying on AI for learning new concepts, as understanding often requires personal engagement and struggle.

[00:12:34] As AI capabilities evolve, staying informed about its strengths and limitations is essential for effective usage.

[00:12:42] Big Technology noted that while consumer adoption of generative AI remains slow, businesses are rapidly integrating this technology to enhance efficiency.

[00:12:51] According to Matt Wood, a former Amazon executive now at PwC, spending on generative AI is almost evenly divided between consumer-facing bots like ChatGPT and APIs designed for enterprise tools, with a larger growth trajectory anticipated for API usage.

[00:13:09] This trend suggests that enterprises are leveraging generative AI to centralize and analyze vast amounts of knowledge, positioning it as a crucial tool for addressing unique business challenges.

[00:13:21] As consumer applications of AI develop, the enterprise's sector investment may provide the necessary funding to drive technology forward.

[00:13:29] And in a detailed report by the AI Now Institute, Bryant Merchant explores the rise of artificial general intelligence and the urgent quest for viable revenue models in the generative AI sector.

[00:13:43] OpenAI, the leading player in this field, reportedly generated an annualized revenue of $3.4 billion in 2024.

[00:13:50] The organization's shifting from a non-profit to a capped profit model to secure necessary investments for its ambitious goals, which have seen notable backers like Microsoft invest $10 billion.

[00:14:02] The report highlights that investment in AI startups surged from $31 billion to $98 billion between 2015 and 2023, with generative AI specifically capturing a significant share of the market.

[00:14:14] However, concerns are emerging about the sustainability of this growth as Goldman Sachs and Sequoia Capital warn that the industry needs to generate $600 billion annually to maintain its investment trajectory.

[00:14:29] As the sector faces challenges related to high operational costs and ongoing legal disputes, the urgency for a clear and profitable business model becomes increasingly critical.

[00:14:40] Why do we care?

[00:14:41] I like asking questions to ponder for these big ideas, so here they are.

[00:14:46] How can you consider the right applications for AI versus the wrong ones for your customers, particularly given this framework?

[00:14:53] With all the coverage of consumer applications of AI, if the real money is in business AI, what does that mean?

[00:15:00] Particularly considering the points raised by the AI Now Institute.

[00:15:04] Ultimately, this has to be profitable for all parts of the supply chain to make sense, including the customer.

[00:15:11] Are we headed there?

[00:15:15] Today's episode is supported by Huntress.

[00:15:18] Huntress.

[00:15:18] You want to focus on your clients and are always looking for ways to get more time.

[00:15:23] Use Huntress's fully managed cybersecurity platform to fight off cyber threats.

[00:15:29] Huntress is more than cybersecurity software for endpoints and identities.

[00:15:33] It's a 24 by 7 security operations center.

[00:15:36] It's security awareness training, community engagement, and dedicated partner support with an average CSAT score of 99.3%.

[00:15:45] Technology can only get you so far.

[00:15:48] Human expertise is what's needed to truly elevate and protect small businesses.

[00:15:53] And you get that with Huntress.

[00:15:55] Secure your clients and help them thrive with the number one rated EDR for S&Bs on G2.

[00:16:01] Visit Huntress.com slash MSP radio to find out more.

[00:16:07] Thanks for listening.

[00:16:09] Today is National Logger Day and International Shareware Day.

[00:16:13] So share a logger?

[00:16:15] Have a question you want answered?

[00:16:16] I take those listener questions, send them in ideally as a voice member or video to question at MSP radio.com.

[00:16:22] I answer listener questions live tomorrow, Wednesday on the live show, 3 p.m. Eastern YouTube and LinkedIn.

[00:16:28] And if you've got a comment or a thought on a story, put it in the comments if you're on YouTube or reach out on LinkedIn if you're listening to the podcast.

[00:16:34] And if you enjoy the show, give it a review and make sure you subscribe or followed on your favorite platform.

[00:16:40] Talk to you again tomorrow.

[00:16:43] The Business of Tech is written and produced by me, Dave Sobel, under ethics guidelines posted at businessof.tech.

[00:16:51] If you like the content, please make sure to hit that like button, follow or subscribe.

[00:16:56] It's free and easy and the best way to support the show and help us grow.

[00:17:01] You can also check out our Patreon where you can join the Business of Tech community at patreon.com slash MSP radio or buy our Why Do We Care merch at businessof.tech.

[00:17:14] Finally, if you're interested in advertising on this show, visit MSP radio.com slash engage.

[00:17:21] Once again, thanks for listening to me.

[00:17:23] I'll talk to you again on our next episode of the Business of Tech.

[00:17:30] Part of the MSP radio network.