Overcoming extreme differences in learning levels amongst students in the same grade is a key challenge for school education in India. Indeed, as ASER 2018 shows, amongst Class III students in Himachal Pradesh, 2.4% couldn\u2019t recognise letters, 10.6% could recognise letters but could not read words, 15.5% could read words but not sentences, 24.1% could read Class I level texts but not higher, and 47.4% could read Standard II level texts. This means the human development potential from the massive jumps in enrollment and narrowing gender gap in classrooms India has seen over the last few decades is being squandered, thanks to policy flaws like the Right to Education (RTE) Act\u2019s \u2018no detention\u2019 provision that was scrapped only in January this year and pedagogical shortcomings. The gap between the level of classroom instruction and the individual learning levels of students is a bottomless rabbit-hole\u2014as the curriculum gets more ambitious with each successive grade, those who were already lagging peers fall farther behind. There has been some action from policymakers on \u201cteaching at the right level\u201d \u2014the Delhi government rolled out its Mission Buniyaad last year to ensure basic reading, writing and arithmetic skills amongst Class III-IX students. But, such interventions are cost-and labour-intensive. Beyond this, as the curriculum gets more complex and learning levels more fractionated, such models may not even be viable. Against such a backdrop, a study by Karthik Muralidharan of the University of California at San Diego, Abhijeet Singh of the Stockholm School of Economics and Alejandro J Ganimian of the NYU Steinhardt School of Culture, Education, and Human Development, shows that greater use of technology in pedagogy could deliver substantial gains. Muralidharan et al study the impact of Mindspark, a device- and connectivity-agnostic piece of technology, that was used to assess learning levels of 619 students from low-income households studying in government-funded schools in Delhi, and address deficiencies. Students in the sample were several levels below their respective grade-appropriate levels. They attended after-school Mindspark centres , where they participated in 90-minutes sessions of instruction, six days a week; the session was divided into 45 minutes of individual self-driven learning on the Mindspark platform and 45 minutes of instructional support from a teaching assistant in groups of 12-15 students. Half of them had been selected, via a lottery, to receive vouchers that made attending the module free. The researchers assessed learning learning levels before and after the four-and-a-half months intervention, and estimated that attending the Mindspark centres for 90 days (or 80% attendance for half a school year) could raise math and Hindi scores with standard deviations of 0.6 and 0.39, respectively. The Mindspark software can determine existing learning levels with greater accuracy\u2014perhaps much more so than even a highly-trained teacher\u2014and dynamically personalise the material being delivered in the instruction process to a student\u2019s level and rate of progress. Also, it delivered near-uniform gains for students irrespective of household socioeconomic status or gender. The researchers estimate Mindspark delivers comparable gains at a fraction of the costs incurred and time consumed in alternative pedagogical approaches, including extra classes. The study offers a sound policy prescription to governments\u2014Delhi, for instance, that has allocated 26% of its budget to education could jump lightyears ahead if it were to complement its Mission Buniyaad with greater adoption of technology in pedagogy.