ShawnRosemarin.com

I can fondly remember as a teenager my uncle talking to me about his graduate studies in computer science and his fascination with neural networks, expert systems and artificial intelligence. At the time it seemed so abstract and ethereal; bandwidth was 1200 baud, CPUs were 33mhz (with the turbo button on) and the closest I could get to anything intelligent was a digital version of the Encyclopedia Britannica and Microsoft Encarta.

Fast forward to 2023 and the AI picture is becoming much clearer. The internet has matured with it we have unlocked a real opportunity to augment human capabilities for the next generation. A quick recap –

• In the 1990’s email, newsgroups and bulletin boards enabled us to share information digitally
• In the late 90’s/Early 2000s evolution in web search allowed us to effectively catalogue and find data across the internet
• In 2009 Wolfram Alpha emerged as the first “computational knowledge engine” connecting the body of knowledge across the internet evolving to power today’s smart assistants like Siri and Alexa
• In the 2010s Netflix and Amazon pioneered the design and integration of predictive analytics into their platforms and driving a high service level to their clients
• In the last 6 months we have seen the beginnings of Generative AI (GAI) across text-to-text, text-to-audio, text-to-video, text-to-code and first looks at projects such as ChatGPT, Dall-E, Jukebox and Codex. Initially, innovation here is being driven by Microsoft, Google and Meta and Amazon but I’ld expect the cast of innovators to expand exponentially over the next 12-18 months.
• In the not so distant future we are likely to see neurotransmitters directly connecting our brains to these systems allowing for human telemetry, intellectual augmentation and autonomic transmission

While it’s clear that we are in the very early innings of GAI I am incredibly intrigued by the potential that this kind of technology can bring to our daily lives. It will likely take us 15-20 years to mature and regulate these learning models but, like the open internet, this new platform presents an incredible space for iterative innovation and one where the next trillion dollar companies are likely to emerge.

The opportunity to integrate GAI across industry and the potential impact to the overall quality of life for the general populations is staggering.

For example –

• In Healthcare we could leverage it to proactively review patient’s pathology, supplementing it with nano-sensors and real-time fitness trackers to identify issues earlier and personalizing treatment
• In Finance we could leverage it to better protect assets, detecting fraud earlier and driving more personalized decisions for consumers of financial products
• In Government we could better serve citizens by streamlining public services, better disseminating information and protecting citizens
• In Education we could vastly richen the learning experience teaching students to leverage GAI as “points of perspective” allowing them to spend less time doing research and more time sharpening their skills around judgement, interpretation and critical thinking
• In Manufacturing we would could identify defects earlier but also optimize the building of goods allowing for lower overall costs and field incident rates
• In Retail we could better serve consumers in evaluating their options but also catering better to their individual needs
• In High Technology we could accelerate software development with generative coding coupled with AI based learning will drive major innovations
• In Transportation we could allow for a more intelligent and catered experience both for not just public transportation but to richen the experience for consumers in the coming era of self-driving cars
• In Media and Entertainment we could see an explosion of generated content driving a new level of personalized consumption for all of us

Of course, along with every opportunity in technology comes a “shadow” with the potential to drive negative consequences. GAI is no different.

• Traditional security models like passwords will no longer be acceptable. Phishing attacks will become more and more difficult to discern from legitimate interactions
• Learning bias will continually be challenged specific to learning libraries and their accuracy, beliefs and completeness
• Intellectual property rights will come under scrutiny as AI models create works that overlap with those under copyright
• Privacy rights will continue to be challenged specific to what information is protected from learning libraries and what is deemed to be public domain

Ill reiterate that we are in the VERY early innings of the GAI evolution. It reminds me very much of the internet back in the late ’90s when we were all moving traditional brochures to HTML for a “digital experience”. From there we evolved to search, e-commerce, audio and video streaming and social media essentially building out today’s internet of people.

As the VP of Customer Engineering at Pure Storage I am looking forward to staying close to this evolution and having many discussions around GAI with our customers. Regardless of the speed of evolution, what we do know is that the volume and velocity of data required to build, train and operate these models will require a robust storage architecture that expands well beyond legacy platforms and the market will be looking at innovators like Pure Storage to solve them.

I encourage you to leave some comments below on any opportunities or threats that you see from this technology to richen the discussion.