So, How is your LLM Today?
Got AI Governance?
AI continues to become the gold rush of technology or a never-before-seen solution. A lot can be said for the world of AI before Deepseek. Deepseek, similar to open-source first generation cloud tools, changed the idea of what AI is really about, lower cost, better modeling.
Why were people so driven to invest in AI well before any form of AI governance, return on investment (ROI), or proof that the user cases could stand the test of time?
How, then, do you measure your LLM? No, not by its height, of course!
LLM: On A More Serious Note
Thanks to #Deepseek, the idea of an AI model being costly changes dramatically. While the #OpenAI and #Gemini all tout their multi-billion investment in #data centers, #nuclear power, and now even coal has a play.
So, if AI is the present, how will these investments build a better future? Is the future more about cost-effective #GPUs and cheaper power, or will open-source LLMs take over from the more seasoned and developed cloud-based hosted models?
Regardless of someone taking up digital estate in one of the 70 data centers near Dulles airport, the idea of open-source means it could be hosted anywhere. The concept of spending huge on AI infrastructure becomes a question of whether it is even needed.
This isn’t the full-time in the history of disruptive technology where overspending, overly complexity of the architecture, and constant pressure for adding new functionality became more important than measuring the cost, efficiency, or relevance.
“You Need More Ports and a Faster Backplane, Honest”
Back in the day, people who hastened moved off traditional network hubs and on to switching infrastructure gained considerable functionality, speed, and scale. Switching provided essential network segmentation, an element of quality-of-service, and a basic level of security.
The cost of moving to a switch was well justified and needed. Hubs were going to end-of-life, and Dot.com websites needed faster backplanes between the web, application, and database layers. Once this progression happened, how did people measure the cost?
The idea of measuring cost-per-port became the standard. However, the cost of the backplane drew attention to the concept of future-proofing the organization, especially with IP telephony, storage area networks, and Wi-Fi networking becoming everyday conversions.
Ultimate application innovation, network consolidations, and the progression towards the cloud became the dominant driving force.
CEOs and the board of directors believed in the notion, “If we are not in the cloud, we are not a market disruptor.”
Regardless of the previous investments made to future-proof the on — premise enterprise network, most of those investments became useless when organizations moved to the cloud.
Hence, the issue is: What will the investments in AI today mean for the future?
AI Becoming The Next Great What Exactly?
Organizations announcing layouts and restructuring banking the efficiency, cost savings, and possible organizational transformation either know something we don’t or are playing with house money.
Even with current investments in cloud-based data lakehouses, LLMs, and everything else AI, without governance frameworks and the commitment of the board of directors to develop a true AI consumption, development, and ethics policy, AI could very well become the boat anchor not seen since the days of Dot.com. Cost overruns, unperforming services, and challenges with retaining the right talent could cause another Dot.com?
What we do we learn from the Dot.com’s? Ask first-generation cloud engineers, architectures, and application developers.
Build it and they will? That is a good question that has yet to be answered.
#AI #AIGovernance #Cybersecurity #CIO #CISO #Application #Development #DeepSeek #AI #ML #Regulatory
