GPT models
Generative Pre-trained Transformer models—AI systems trained on large text datasets that generate human-like language in response to prompts.
As defined in Tyler Cowen on the Risks and Impact of Artificial Intelligence
Where it comes up · 1
Most important claims · 5 of 5Search all claims for “GPT models” →
- One of Cowen's predictions is that GPT models will raise the relative and absolute wages of carpenters; people who produce routine word-based content should seriously rethink their careers, while people with synthetic abilities who can physically make things happen—good gardeners, carpenters, lab builders of fusion reactors—are already becoming much more valuable.Tyler Cowen on the Risks and Impact of Artificial Intelligence
- Because GPT models can produce content that is very hard to identify as machine-generated when prompted well, the age of homework is over; rote learning will be worth much less in labor markets, and educational systems—especially East Asian ones that stress homework heavily—will have to reorganize fundamentally, a hard adjustment given how slow educational institutions are to adapt.Tyler Cowen on the Risks and Impact of Artificial Intelligence
- In the Austrian Economics worldview, decentralized and often inarticulable (tacit) knowledge as outlined by Michael Polanyi is critically important to almost everything that happens; GPT models do not directly access or train on that knowledge—they only digest a version of its outputs—so although they have cognitive oomph, AGI is not well-defined and intelligence's multi-dimensionality (including realizing intelligence through physical action, as LeBron James does) means timelines like 'AGI is X years away' do not make sense.Tyler Cowen on the Risks and Impact of Artificial Intelligence
- In the short-to-medium term, much routine back-office work will be done by GPT models, and organizations will find new ways of unsiloing, organizing, storing, and accessing their information—like Star Trek's Spock querying a computer conversationally—making many institutions much more efficient, though not boosting GDP growth to anything like 30% a year.Tyler Cowen on the Risks and Impact of Artificial Intelligence
- AI will disrupt many of us—often in good ways, sometimes bad—so everyone should reevaluate their professional lives now and look for things they can do to make this a better future rather than a worse one, urging against complacency; Cowen even doubts he will write another book since people will play with their GPT models rather than read it, and plans to give more talks instead.Tyler Cowen on the Risks and Impact of Artificial Intelligence