This is amusing:
One of Rupert Murdoch’s top lieutenants has warned that AI threatens a “tsunami” of job losses and could crush readers under a weight of “maggot-ridden mind mould”.
Thomson said that AI will lead to a “tsunami” of job losses. “From 2008 to 2020, 57 per cent of newsroom jobs in the United States have been lost,” he said.
…While warnings about AI job losses are not limited to newsrooms, Thomson warned that there was a greater societal risk: that we will become deluged by a stream of AI-generated “rubbish”.
Err, I can’t see how this differs from today’s iMSM. It has already lost any insight it ever had. It has turned ideologically activist across nearly all mastheads. It long ago sacrificed investigative firepower for access journalism.
These days, it is all just one big whine from the interests running the show and the individuals marginalised by it. There is no context or cause and effect.
The iMSM is already maggot-ridden mind mould. I read a piece worth my time perhaps once per quarter—maybe less.
Doubtless, AI rubbish will flow. And because it is cheap to produce, there will be more of it. But, the better stuff will be more rational and independent than most journalists, and the media market will still offer plenty of choice.
As a result, trusted mastheads and content will also rise in value. Those that can offer something that neither AI nor teen press-release producers can. Genuine insight. That increased value should generate investment in quality to grow eyeballs.
I can see AI being quite good for journalism. Increased profits from shedding dead wood can be recycled into developing humans with genuine insight and more investigative pieces.
If I am wrong, and AI produces more immigration and real estate cheerleading, then at least consumer prices will fall for the same rubbish they are forced to read today.
Bring AI on. Journalism can’t get worse than it already is.
Or am I being too cavalier?
We explore the ability of Large Language Models (LLMs) to produce conditional inflation forecasts during the 2019-2023 period. We use a leading LLM (Google AI’s PaLM) to produce distributions of conditional forecasts at different horizons and compare these forecasts to those of a leading source, the Survey of Professional Forecasters (SPF). We find that LLM forecasts generate lower mean-squared errors overall in most years, and at almost all horizons. LLM forecasts exhibit slower reversion to the 2% inflation anchor. We argue that this method of generating forecasts is inexpensive and can be applied to other time series.
Economics can do with the shakedown as well…
