The daily SignalSignal · Ep 124 · November 20, 2026

Cheap Models, Fragile AI Advantage

OpenAI and Google cut capable model pricing again, and that changes more than your A P I bill. If your AI edge depends on one favorite model, it is weaker than you think. The smart move now is routing, portability, and owning the workflow around the model.

Listen now · Ep 1240:00 / 4:47
Your question of the day

If capable AI gets cheaper again next quarter, what part of my workflow, customer context, or data would still give me a real edge?

Stuck? Tap a starting point and make it yours:
A free account unlocks: space to reflect on what you picked up each week
Your answer is saved to your private log the moment you sign up — free.

Pro adds the new skill series — like Excel + AI — dripped in with your daily Signal.

OpenAIGoogle GeminiAI pricingworkflow strategymodel routing
Transcript· the complete episode, word for word

Morning. Damian built an AI version of himself for the daily briefing. Useful little machine… especially before the real one reaches full operating voltage. DayLift Signal. AI-curated. Five minutes.

Capable AI just got CHEAP again. That is not exciting news — it is dangerous news for lazy strategy. I read through the Friday pile this morning… most of it was model chest-thumping. This is the one shift that actually hits your work.

OpenAI cut Astra pricing again, and Google launched Gemini four Neon at three dollars per million input tokens and fifteen dollars per million output tokens. Two big vendors, same direction, same week. The verdict is simple: general-purpose model power is turning into a commodity faster than most teams are planning for… and that means your edge cannot live in the model.

Team leads and managers — this hits your operating math first. If your team uses AI for drafting, extraction, summaries, research, or support prep, the old cost assumptions are already wrong. Workflows that looked too expensive last month may now be worth rolling out. Other workflows may prove they never needed a premium model at all. Owners and decision-makers — this is the bigger story for you. Reopen any build-versus-buy call from this quarter. If your AI bet depends on one expensive model staying special, that bet is getting weaker by the week. Individual operators and solo professionals — worth knowing, but this is NOT really your day unless client delivery already runs through an A P I at volume. Smart move: make the workflow PORTABLE, then let models compete for the job.

Here is the lever. Team leads, this is your move first — owners should ask for the numbers. Export one week of AI tasks from ChatGPT, Claude, Gemini, or your A P I logs. Label each task routine, ambiguous, or high-risk. Then test twenty real examples through a cheaper model like Sol or Gemini four Neon, and escalate only the weak outputs to a stronger model. Start with drafting, classification, extraction, and summaries. Measure accepted quality, review time, and total spend. If you handle sensitive customer, employee, or financial data, keep it inside approved business agreements only.

Here is my honest take… I think a lot of people got way too emotionally attached to one favorite model. Every few weeks the leaderboard shifts, the pricing shifts, the vibe shifts. REAL strategy is not loyalty to a model. It is owning the workflow, the customer context, and the tests that tell you whether the output is good enough to trust.

This is the trap I keep seeing in growing companies. A team spends months polishing prompts, wiring one model into a slick demo, and calling that innovation. Then the price drops, quality changes, or another platform ships the same thing natively. You're still calling model choice strategy when the real asset is the workflow around it. Of course that stings… because the demo may work. Better pattern: prove the bottleneck first, keep human checkpoints, structure the inputs, and make the model replaceable from day one.

So here is the question. If capable AI gets cheaper again next quarter, what part of your workflow, customer context, or data would still give you a real edge?

Get the next one automatically

This is one of the daily Signals. Sign up free and tomorrow's lands in your inbox — plus the question, the prompt of the day, and the Academy when you want to go deeper.

[matter-of-fact] DayLift Signal. AI-curated. Five minutes. [short pause]

This episode is read by a disclosed AI clone of the founder's voice. Content created with AI assistance and reviewed by a human. How this is made

More recent Signals

Ep 330Accounting AI Permissions Just Became StrategyEp 123AI Turns Webinars Into Revenue DataEp 329ChatGPT Enters Core Accounting Work