deniz.in

Markets

Weather

Loading weather

· via Hacker News – Front Page (native)

Alpha School's own data shows most students miss its twice-as-fast learning goal

Only 42% of students at the $75,000-a-year AI-driven Alpha School met its goal of learning twice as fast in math and reading, internal data shows, even as the network pushes its software into public districts.

Alpha School's own data shows most students miss its twice-as-fast learning goal

Alpha School, the private network charging families up to $75,000 a year for an AI-driven education, is falling short of the central promise of its own model. Internal figures first reported by The New York Times and reviewed by Fortune show that just 42% of Alpha students met the network's target of learning twice as fast as their peers in both math and reading.

Math is the bigger gap

Roughly half of kindergarten-through-eighth-grade students at Alpha's private campuses missed the 2x growth goal in math, while about one in five fell short of their reading targets, according to the Times' review of the network's data.

Students who started behind fared worst. About a fifth of private-campus students scored at or below the 50th percentile nationally in math last fall, and only 37% of that group reached the 2x growth goal by winter testing.

How the model is supposed to work

Alpha employs no traditional teachers. Instead, staff called guides focus on keeping students motivated; they do not plan lessons or grade work, as Fortune reported in September. The academic core is a two-hour daily session with TimeBack, an app Alpha markets as an AI tutor. Cofounder MacKenzie Price says the app personalizes and speeds up learning for each child, freeing the rest of the day for workshops on skills such as public speaking.

The network has grown quickly, though the exact size is unclear: a spokesperson told Fortune in September that Alpha had more than 1,200 students, while the Times put the figure at about 1,500 across roughly 50 campuses.

The pitch has drawn prominent support. U.S. Education Secretary Linda McMahon toured Alpha's Austin campus in September 2025 and described it as the most exciting development she had seen in education in years, according to Fortune. LinkedIn cofounder Reid Hoffman and hedge fund manager Bill Ackman have publicly backed Price, and even the X account of imprisoned Theranos founder Elizabeth Holmes has praised her.

Strain behind the two-hour day

Reporting from those closest to the model complicates the two-hour premise. Guides and parents told Wired that students regularly worked late into the night to hit the school's goals, and one parent of a 13-year-old reading at an 11th-grade level said reaching those targets came with stress and anxiety.

Former employees told 404 Media that Alpha's AI generated faulty lesson plans — internal documents reportedly acknowledged these could do more harm than good — and that the company scraped other online courses without permission.

A harder test in public schools

The internal data lands as Alpha expands beyond its private campuses. More than 2,600 public school students in Texas and South Carolina now use its platform, according to company documents reviewed by the Times. Public schools are a tougher environment: districts cannot remove struggling students and do not use the prize incentives Alpha deploys at its private sites.

Houston Independent School District, Texas's largest, is piloting the software at two schools. Superintendent Mike Miles told the Times that students spent more time on Alpha's apps than planned, cutting into other activities, though he still wants AI in the district's toolkit.

The broader research picture

Alpha's results arrive alongside growing evidence that how students use AI matters as much as whether they use it. A study of 26,811 secondary students in central China found homework scores rose 18% after students adopted generative AI tools, while closed-book exam scores fell 20%. The OECD's latest PISA results showed daily AI users scoring lower in science than nonusers, with moderate users performing about the same.

Professors have also told Fortune that students who lean on AI summaries struggle with sustained reading, with Notre Dame's Timothy O'Malley and Pepperdine's Jessica Hooten Wilson describing students unable to work through long texts.

Why it matters

Alpha is one of the most visible tests of the claim that AI tutors can dramatically compress learning time. Its own internal numbers — not those of outside critics — now show most students missing the network's headline goal, with math outcomes weakest and struggling students benefiting least. That matters beyond the families paying $75,000 a year, because public districts are already adopting the same software under tighter constraints. As AI-education spending scales on optimistic promises, this data is a concrete signal that personalized AI tutoring has not yet delivered the learning gains its marketing implies, and that results in selective private settings may not transfer to public classrooms.

  • #ai-education
  • #edtech
  • #ai-tutors
  • #education-policy
  • #private-schools