AIQuiet 12d · day 12
Dan Shipper: New Foundation Model Produces Probabilities, Not Words
Early tester signals a foundation model that outputs probabilities rather than text could become indispensable within 6-12 months.
What to know
- A new foundation model in testing produces probabilities rather than text, enabling use as a decision-maker or judge.
- Early tester signals the model could become indispensable within 6-12 months based on one week of evaluation.
- Limited public information available; development details and model name not disclosed in available coverage.
Dan Shipper Early tester at Every
How it unfolded 1 development · click the chart to see its coverage posts
Sep 16Sep 17Sep 18Sep 19Sep 20Sep 21Sep 22Sep 23Sep 24Sep 25yesterdaynow · 11:07 PM ET
-
1
Dan Shipper reports early testing of probability-based foundation model
Shipper posted that he and colleagues at Every have been testing a new foundation model for approximately a week, noting it produces probabilities rather than words as output, enabling it to function as a decision-maker or judge.
“we almost never test new foundation models but we've been testing this for ~a week @every and it's pretty wild. the kind of things that will be obviously indispensible in 6-12 months…”
— Dan Shipper -
we almost never test new foundation models but we've been testing this for ~a week @every and it's pretty wild. the kind of things that will be obviously indispensible in 6-12 months it doesn't produce words as output, it produces probabilities. so it can efficiently act as a
-