On the heels of another impressive AWS re:Invent in Las Vegas, it feels as if we all got a software update! For some of our team, it was their first re:Invent, which is a feat in itself. As an active cloud architect myself, I’ve attended this monumental affair for the last decade as a speaker, partner, AWS Hero, and user. Each time, I leave refreshed with new knowledge, friendships, and inspiration.
Having enjoyed our time together taking in the wide selection of technical information, attractions in Sin City, and most importantly, meeting fellow cloud geeks, there’s a lot that came out of the event.
As aiOla’s CTO and an AWS ML hero, these are my main takeaways.
TLDR;
In just four words: “Do more with less.” AWS’ main focus is on helping large enterprise workloads and those looking to modernize their systems at lower costs with improved security, and greater flexibility.
AWS has rushed generative AI products such as Amazon Q, which will need a lot of work from partners to make it practical for most companies to use. (Continue reading for more product reveals).
For us, at aiOla, we’re riding the same wave. Our solution is geared towards businesses that want to “do more with less” and grants them the ability to capture better data and boost efficiency in their existing internal workflows, all through the use of speech-powered technology.
aiOla’s Evolution in re:Invent
aiOla’s business is evolving quickly, and with it, so is the technology stack and the cloud infrastructure. In previous years, we invested a lot into cloud agnostic infrastructure with Kubernetes and Terraform. At the beginning of this year, we simplified the product into a SaaS offering that can be customized in just days for new customers and start providing value in a couple of weeks.
From the technical perspective, we had to deepen the team knowledge in AWS best practices and managed services. re:Invent is packed with technical content for all levels of experience and personas, and so, we decided to send a significant number of team members to participate in the event.
Sadly, many plans quickly changed due to the horrific attack on Israel on October 7th. Many of our team members who were scheduled to attend the event in Las Vegas were recruited or volunteered to reserve duty in the IDF, or are spending long nights and weekends in civil war rooms that were created to solve the many challenges the war created.
As such, most of the original team members had to cancel their trip. But, as the conference drew near, we managed to find new team members to travel in these trying times, committing to our promise that “we deliver NO MATTER WHAT”.
Hey, Enterprises: AWS Wants Your Workloads
This year, the top-line message out of the conference was dedicated to the enterprise workload.
While the focus on enterprises is not new, during this year’s re:invent 2023, the messages were even clearer with announcements of:
- New support for DB2 in RDS
- AI assistance in upgrading Java code to latest versions
- CloudWatch connection to multi-cloud and on-premise applications
These updates might seem random and unrelated; however, if you put them on top of the many security features, data migration, and even mainframe modernization, AWS now covers almost every legacy system that exists in enterprise companies.
At aiOla, we also see the growing maturity of enterprise customers to adopt cloud-based systems. A few years ago, being a cloud product required longer and tedious security reviews and delayed some of our early implementations. Now, these blockers are mostly gone, making it easier and more preferred for a cloud product to make it through the IT gates with welcoming arms.
A Low-Cost Solution for All
Enterprises can’t change their IT overnight to support micro-services, event-driven applications, or other cloud native architectures. But, they are happy to save costs with migrations of their current applications to the cloud with cheaper compute (Graviton 4 for databases, and Inferentia 2 for AI inference, for example), built-in high availability and backups.
In his keynote, Amazon CTO Werner Vogels spent the majority of his time talking about “the frugal architect,” and the ways to save on costs by better utilizing the resources that are available.
From aiOla’s perspective, we agree that AI is not here to replace people, as it is impossible to replace human intelligence and experience beyond simple tasks. That’s why our product leverages AI to augment employees’ capabilities to be more productive and work more safely during critical workflows, such as inspections and other reporting activities.
Innovation to “do more with less” is easier to understand and implement over a complete “AI transformation,” which tends to fail.
The Need for Technology Partners is Real
As expected, generative AI and AI in general was front and center in the leadership keynotes.
In the heroes community, there were bets (afterall, it is Las Vegas) on the number of times Adam Selipsky would say AI in his keynote -– with some bets reaching over 300. Eventually, we calculated the rate of AI mentions, which was more than once per minute.
The main AI product announcement was Amazon Q, which uses generative AI technologies and LLM models to index internal enterprise documents and answer questions based on these documents.
AWS also announced Q Architect, which is built on top of Amazon Q and is based on “17 years of experience.” The Q architect depicts the potential of such AI-based products. At the same time, it also shows the long way it still must go and the technical help that’s required to make it useful.
AWS had to limit the service to focus on architecture from the greater vision of “Ask AWS” to support other personas. Even with the mass amount of effort needed to make it work, many of the answers provided by the service (scroll down in this summary for examples) are inaccurate, outdated, and lack best practices.
At aiOla, as a technical partner for large enterprises with AI-based products, this reality gives us the confidence to continue in our mission in making AI successful for our customers. The secret sauce of deriving the real value of AI is driven by our ability to customize our AI models based on the specific needs of the many large enterprises that we serve.
This End is Just the Beginning
re:Invent is a huge technology conference, and so, it is impossible to summarize it all.
My goal here was to share the overarching highlights in an effort to help other IT professionals make sense of the many service announcements and hundreds of hours of content (some of which is available online).
