Rethinking AI in the Academe: Hardware Depreciates, Skills Scale

I saw this presentation posted on a page for the #CHEDRAISE event earlier today, and there was a strong push for schools to acquire heavy on-premise AI infrastructure for student learning and research.

While the goal of bringing AI to education is highly commendable, as someone who builds and integrates these systems, I have to point out that buying physical AI servers is a costly trap for most institutions. Here is the reality of on-premise AI:

1. 𝐌𝐚𝐬𝐬𝐢𝐯𝐞 𝐂𝐚𝐩𝐄𝐱: You're looking at $30,000+ per unit upfront.
2. 𝐑𝐚𝐩𝐢𝐝 𝐎𝐛𝐬𝐨𝐥𝐞𝐬𝐜𝐞𝐧𝐜𝐞: Frontier AI models evolve at lightning speed. The expensive machine you buy today likely won't have the computing power or parameters to handle next year's use cases.
3. 𝐇𝐢𝐝𝐝𝐞𝐧 𝐎𝐩𝐄𝐱: Training even small models requires running machines 24/7. This doesn't even factor in the massive power bills, specialized cooling systems, and dedicated IT maintenance.

𝐓𝐡𝐞 𝐑𝐞𝐚𝐥𝐢𝐭𝐲 𝐂𝐡𝐞𝐜𝐤 𝐟𝐨𝐫 𝐭𝐡𝐞 𝐏𝐡𝐢𝐥𝐢𝐩𝐩𝐢𝐧𝐞𝐬:

We need to stop trying to "reinvent the wheel." In the Philippines, we simply do not have the financial, technological, or human resources to compete in building frontier models from scratch.

Instead, our focus should be on contextualized use cases. Rather than spending millions on hardware to train models that others have already perfected, we should be investing in applying these tools to solve our unique local challenges. There are many better, scalable, and cost-effective cloud alternatives available today that allow us to do exactly that.

Schools shouldn't be pouring their limited budgets into rapidly depreciating hardware. Instead, we need to invest in our PEOPLE.

Seminars are a good start, but true technological adoption happens when your team rolls up their sleeves and 𝐠𝐞𝐭𝐬 𝐭𝐡𝐞𝐢𝐫 "𝐡𝐚𝐧𝐝𝐬 𝐝𝐢𝐫𝐭𝐲" 𝐛𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐩𝐫𝐚𝐜𝐭𝐢𝐜𝐚𝐥 𝐬𝐨𝐥𝐮𝐭𝐢𝐨𝐧𝐬. Prioritize building capable teams over buying heavy metal (before it's too late!).

People who know what they are doing will help organizations determine the proper use cases to focus on, rather than fall victim to the 𝐬𝐡𝐢𝐧𝐲-𝐨𝐛𝐣𝐞𝐜𝐭 𝐬𝐲𝐧𝐝𝐫𝐨𝐦𝐞, as most organizations are doing right now, whether in the academe or industry.

Would love to hear from other educators and IT professionals. How is your organization handling the push for AI?