DeepSeek released V4.1 Flash on September 10, the latest in a line of fast, low-cost models from the Chinese lab that has repeatedly pushed the price of capable AI downward. "Flash"-class models trade a little peak capability for much lower cost and quicker responses, which suits high-volume, everyday tasks.
Each new low-cost entrant makes it harder for any provider to charge a large premium for ordinary work, benefiting anyone who buys AI by the request.
What it means: the cost of AI good enough for routine business tasks keeps falling, so a job that looked too expensive to automate a year ago may be affordable now. It is worth re-pricing automation ideas that were shelved on cost, because the numbers have likely changed.