Key elements of Friedberg’s prediction have not materialized by the forecast horizon (mid‑2023) or even by late 2025:

1. “Many” new drugs on the market within a couple of years
   • As of mid‑2023, the only new disease‑modifying monoclonal antibodies approved for Alzheimer’s in the U.S. were aducanumab (Aduhelm, 2021) and lecanemab (Leqembi, accelerated approval Jan 6, 2023; traditional approval July 6, 2023).(alz.org)
   • The next major antibody, donanemab (Kisunla), was not FDA‑approved until July 2024.(depts.washington.edu)
   • Including older symptomatic drugs (donepezil, rivastigmine, galantamine, memantine) doesn’t change that there were not “ten things approved” in the couple of years after June 2021—only one controversial new drug (Aduhelm) plus Leqembi right at the end of that window.

2. “Everyone’s going to be taking an Alzheimer’s drug in a couple of years” (near‑universal use)
   • Real‑world data show very low uptake of the new antibodies. A JAMA Network Open study of Medicare fee‑for‑service beneficiaries found 1,725 lecanemab users among 842,192 patients with diagnosed Alzheimer’s or MCI between July 2023 and March 2024—uptake well under 1%.(medicalxpress.com)
   • Earlier U.S. work on symptomatic drugs (cholinesterase inhibitors and memantine) found that only about 33–56% of dementia patients ever received an approved antidementia drug, not “effectively everyone.”(pubmed.ncbi.nlm.nih.gov)
   • Globally, care gaps are huge: a Yale‑linked global study reports that around one in five people with dementia receive no care at all, even counting non‑drug support, implying far less than universal medication use.(ysph.yale.edu)
   → The “everyone is taking an Alzheimer’s drug” part of the prediction is clearly false.

3. “Maybe 10% of people get cured, 90% don’t”
   • No approved Alzheimer’s drug is described by regulators or major medical organizations as a cure. The Alzheimer’s Association explicitly calls Leqembi “not a cure,” saying it only slows progression in early disease.(prnewswire.com)
   • News and medical coverage of Leqembi and similar drugs consistently state they “modestly” or “moderately” slow cognitive decline, not that they cure any subset of patients.(cbsnews.com)
   • There is no evidence that ~10% of all treated Alzheimer’s patients are cured in the ordinary sense (disease eradicated or permanently halted).
   → The cure‑rate claim is not supported by current data.

4. Per‑patient cost “on the order of millions of dollars”
   • List prices for the key disease‑modifying antibodies are high but orders of magnitude below “millions” per patient:

   • Leqembi (lecanemab): about $26,500 per year in the U.S.(prnewswire.com)
   • Donanemab (Kisunla): roughly $32,000 per year.(news.northeastern.edu)
   • Aducanumab (Aduhelm): initially $56,000 per year, later cut roughly in half before discontinuation.(grandviewresearch.com)
     • Analyses that include required MRI/PET imaging and monitoring estimate total per‑patient annual costs around $33,000–$34,000 for Leqembi, not millions.(kffhealthnews.org)
     • Over many years of continuous treatment, some patients could approach low six‑figure totals, but “millions of dollars per person” is well above current real‑world or modeled course‑of‑treatment costs.

5. “Very high system‑wide healthcare spending”
   • On this narrower point, projections do show substantial potential spending: JAMA Internal Medicine modeling estimates that broad Medicare coverage of Leqembi could add $2–5 billion per year in Medicare costs alone.(pmc.ncbi.nlm.nih.gov)
   • However, because uptake has remained limited so far (e.g., a few thousand users relative to millions of U.S. patients with Alzheimer’s), actual realized spending through 2025 is far below the worst‑case projections.

Overall assessment
The central quantitative and qualitative claims—many new drugs within ~2 years, near‑universal treatment, ~10% cures, and per‑patient costs in the millions—are all inconsistent with what has actually happened by mid‑2023 or even by late 2025. While the general direction (new expensive drugs creating significant aggregate spending with modest benefit) has some truth, the forecast was far too aggressive in scale, timing, uptake, and clinical effect.

Therefore, the prediction is best classified as wrong.