Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
68 61 75 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
42 25 60 -
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
Insufficient record: 0 promises resolved, 5 needed for a score.
All types combined: 65 (likely 58 to 71, high confidence), from 78 claims: 69 factual claims, 9 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Ray Kurzweil's claims
18 claims
“And then AlphaFold 2 in 2022 went through 200 million proteins and got the shape exactly correct, and that’s going to help come up with new medications.”
Ray Kurzweil, futurist, interviewee 9 July 2024 'The Singularity is Nearer,' Says Futurist Ray Kurzweil (Science Friday)
AlphaFold 2 in 2022 went through 200 million proteins and got their shapes exactly correct, which will help produce new medications.
Mostly false False in substance, right in a detail.
What would count as true or false
Three elements: (1) AlphaFold 2 in 2022 processed ~200 million proteins; (2) it got their shapes 'exactly correct'; (3) this will help produce new medications. True if DeepMind records confirm the ~200 million figure and 2022 timing, the predictions are described as highly/exactly accurate, and the medical-application claim is supported. Note the released database (~200M structures) was published mid-2022. Mostly true if the 200M/2022 facts hold but 'exactly correct' is a reasonable lay overstatement of high (not literally perfect) accuracy. Mixed if roughly half the elements hold. Mostly false if the core numbers are right but 'exactly correct' materially distorts AlphaFold's actual accuracy (predictions are confidence-scored, not verified exact). False if the 200M figure or 2022 timing is wrong.
Evidence
- AlphaFold predicts structure of almost every catalogued protein known to science | EMBL EMBL · 28 July 2022
“The database is being expanded by approximately 200 times, from nearly 1 million protein structures to over 200 million, covering almost every organism on Earth that has had its genome sequenced.”
- AlphaFold predictions are valuable hypotheses and ... Nature Methods (Springer Nature) · 30 November 2023
“Our results also show that even very high-confidence AlphaFold predictions differ from corresponding models deposited in the PDB by about twice as much as pairs of high-resolution structures in the PDB that were crystallized in different space groups”
- AlphaFold predictions are valuable hypotheses and ... Nature Methods (Springer Nature) · 30 November 2023
“AlphaFold predictions have already been made available for 200 million individual protein sequences to further drug discovery and protein engineering and to understand biology”
- Researchers Assess AlphaFold Model Accuracy - Biosciences Area Lawrence Berkeley National Laboratory, Biosciences Area · 23 January 2024
“which makes those parts of the predicted models unusable for detailed analyses such as those needed for drug discovery.”
Reasoning
E1, dated July 2022, confirms that the AlphaFold database grew to more than 200 million predicted protein structures that year, so the number and the timing hold. E3 says the predictions were released to support drug discovery, which backs the medical-use part in general terms. However, E2, a peer-reviewed study, found that even very high-confidence AlphaFold predictions differ from the experimentally determined structures in the Protein Data Bank (PDB) by about twice as much as two independent experimental structures of the same protein differ from each other. E4 adds that parts of some predicted models are unusable for detailed work such as drug discovery. With the core facts right but "exactly correct" misstating what are confidence-scored predictions, the criteria point to mostly false.
Cite this claim
Ray Kurzweil, “AlphaFold 2 in 2022 went through 200 million proteins and got their shapes exactly correct, which will help produce new medications.” (said 9 July 2024). Verdict: mostly false. Reliability API, claim c_c28fe75a78c1, method 0.1.0+28da144b. https://epistacklabs.com/claims/c_c28fe75a78c1
Claims without a verdict
Is a verdict here wrong, or do you speak for Ray Kurzweil? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.