Now that the holiday season is in full swing, and the long trip home from Las Vegas has given us an opportunity to decompress from the week-long madness, it’s time to try to assemble a coherent picture of what we saw and heard at AWS re:Invent 2023.
Our experience at the big show
Our booth was busy with visitors, in no small part due to the very enthusiastic SnapLogic people staffing it. I mean, check out this pitch! On showcase was our LLM-powered generative intelligence functionality, SnapGPT (powered by AWS Bedrock), which, almost uniquely on the floor at re:Invent, is already shipping and in production in customers’ environments.
We shared a peek at our forthcoming no-code, GenAI-powered solution for application integration and automation, which you can learn more about from this recorded webinar. And of course we had all the existing integrations with any number of AWS services: PrivateLink, S3, SNS, SQS, Redshift, DynamoDB, and more.
The keynotes and sessions were full of excitement. In addition to those external speakers, our CTO, Jeremiah Stone, joined executive panels with exciting companies such as Anthropic and the host AWS itself. Both were lively discussions, centering on the potential of generative intelligence and how best to leverage it.
We were thrilled to be joined in Las Vegas by two thriving companies and SnapLogic customers who shared engaging presentations about how SnapLogic has enabled them to their digital transformation goals.
Siemens Healthineers, a leading global MedTech company, had a vision and embarked on an ambitious goal to empower citizen integrators to create their own data flows across the enterprise. Data flows covered a spectrum of enterprise applications and business processes such as customer services, CRM, finance, and PLM.
In their talk, Siemens Healthineers showed how to maximize the scale of enterprise integration deployment across IT, business groups, and thousands of users while minimizing DevOps efforts. All this was accomplished with the speed and ease of use of the SnapLogic no-code Intelligent Integration platform combined with the breadth and power of AWS infrastructure and cloud services.
Lumeris, a value-based care enablement partner to health systems and physician groups, navigated attendees through the evolution of an initiative engineered to rejuvenate and fortify Lumeris’s AI-powered data and analytics capabilities with SnapLogic. Now Lumeris empowers data and analytics teams to create custom datasets in minutes to accelerate response times and informed, data-backed clinical point-of-care decision-making.
And for those following along at home, I didn’t quite beat my average step count from last time — but these are pretty respectable totals nevertheless!
The takeaway: everyone is betting on Generative Intelligence
As expected, generative AI was absolutely everywhere at the conference: on every vendor’s booth, in every conversation, and in every presentation — although Adam Selipsky did make it an admirable twenty minutes into his opening keynote before raising the topic. I posted my live reactions to his talk in my comments over on LinkedIn.
Despite his restraint, though, it did seem at times during the week of re:Invent that “AI with everything” might be a good one-sentence summary.
In the run-up to the show, a lot of presentations presumably had to be rewritten in a hurry several times over in response to the OpenAI drama that unfolded a scant week before everything kicked off in Las Vegas. On the other hand, it’s important not to over-index on the news headlines.
Between the creation of AWS Bedrock and their investment in Anthropic, AWS was already diversifying long before re:Invent. At SnapLogic, we see many of the same factors at play, which is why we offer our users a choice of models through our partnership with AWS Bedrock.
The Amazon Q announcement, on the other hand, fell a little flat. I will qualify that statement right away by pointing out that the audience at re:Invent is primarily made up of developers, and ones that are pretty knowledgeable about AWS at that, so they are perhaps not the target audience for Q. Its initial functionality is focused on helping navigate the AWS console, and re:Invent attendees are almost by definition pretty expert in that arcane art.
GenAI in action
When it came to more concrete discussions of how to take AI into production, I appreciated Dr. Swami Sivasubramanian’s emphasis on the importance of data. Again, as I was commenting over on LinkedIn, it’s important not to get too carried away, here. As with previous waves of AI, it’s not as simple as just pointing the AI at a huge heap of unsorted data and waiting for magic to happen. The data is indeed a prerequisite, but just getting access to it from where it currently lives, dispersed across the entire enterprise IT landscape, can be a challenge, before we even get to questions of data quality and governance.
This is why a data fabric approach, enabled by SnapLogic, is so important to data-driven projects, whether they use AI or not.
“The Frugal Architect”
I always look forward to Werner Vogels’ closing keynote, with his combination of theatrics and solid advice born from experience. My live reactions are also over on LinkedIn. His theme of “frugal architecture,” equating cost with sustainability, is perfectly timed, as this is a conversation I am having a lot with SnapLogic customers.
Idle infrastructure consumes resources, both financial and ecological, so anything that we can do to optimize that usage is a win on several different levels. Whether this is as simple as any of the following, the benefits compound over time:
- Adopting ELT patterns in place of ETL
- Pushing down complex transformation to on-demand cloud data warehouse infrastructure
- A more complex re-architecting of API integration
Improving the odds: SnapLogic’s generative integration solution
It was surprisingly hard to identify a single overall theme of this year’s re:Invent, beyond the obvious one of generative intelligence. Perhaps wisely, AWS was reluctant to lean all the way into making that the official theme of the conference — but that reluctance left a blank space with no other obvious candidate to fill it. Attendees were certainly eager to discuss the topic, as long as there was a concrete application for it. This is a sign of a maturing market: it’s no longer about the technology, but what we can do with it.
SnapGPT (as well as some of the even newer SnapLabs prototypes) fit that bill. I saw several visitors to our booth ask to take over the keyboard and try their own prompts. It’s sometimes hard to recall that the current boom in generative AI is barely a year old at this point!
The most astute observation I heard all week was that the trajectory of generative intelligence mirrors that of the cloud, with its separation into infrastructure, platform, and applications. Much as we had to deal with a certain amount of cloud-washing in the early days of cloud, there is plenty of opportunistic talk of AI right now — but it’s already shaking out, with concrete applications and even live customer success stories emerging. As AWS people might say, it’s still day one. It will be interesting to see how things progress in 2024.
On that note, I had a ton of great conversations at the show, and now I need to go follow up on them and try to make our ideas concrete. Watch this space!