| The next article initially appeared on Medium and is being republished right here with the creator’s permission. |
Don’t get me fallacious, I’m up all evening utilizing these instruments.
However I additionally sense we’re heading for an costly hangover. The opposite day, a colleague instructed me a couple of new proposal to route one million paperwork a day by way of a system that identifies and removes Social Safety numbers.
I joked that this was going to be a “million-dollar common expression.”
Run the mathematics on the “naïve” implementation with full GPT-5 and it’s eye-watering: One million messages a day at ~50K characters every works out to round 12.5 billion tokens each day, or $15,000 a day at present pricing. That’s almost $6 million a 12 months to verify for Social Safety numbers. Even in the event you migrate to GPT-5 Nano, you continue to spend about $230,000 a 12 months.
That’s a hit. You “saved” $5.77 million a 12 months…
How about working this code for one million paperwork a day? How a lot would this price:
import re; s = re.sub(r”bd{3}[- ]?d{2}[- ]?d{4}b”, “[REDACTED]”, s)
A plain outdated EC2 occasion may deal with this… A single EC2 occasion—one thing like an m1.small at 30 bucks a month—may churn by way of the identical workload with a regex and price you just a few hundred {dollars} a 12 months.
Which signifies that in observe, firms will likely be calling individuals like me in a 12 months saying, “We’re burning one million {dollars} to do one thing that ought to price a fraction of that—are you able to repair it?”
From $15,000/day to $0.96/day—I do assume we’re about to see numerous firms notice {that a} considering mannequin linked to an MCP server is far more costly than simply paying somebody to write down a bash script. Beginning now, you’ll be capable of make a profession out of un-LLM-ifying functions.
