Ben Stern is “the Manager of Education Partnerships at TeachBoost and an advisor to Ponder. Formerly a history teacher and technology integrator….” This appeared in EdSurge September 11, 2014.

Most recently, Andreessen and Horowitz suggest healthcare, education, and government are poised to be eaten. They may be right about healthcare. Government, they admit, is a longer way off. Education may be more imminent. Andreessen explains,

Technology is not driving down costs in…education, but it should…[Access is] the critical thing. We need to get every kid on the planet access to what we consider today to be a top Ivy League education. The only way to do that is to apply technology.

Ostensibly supporting this thesis is the growing edtech industry, in which I work. $452 million of VC money in 2013 is no small potatoes. Andreessen and Horowitz must be on to something, right?

No – at least, not as they frame the issue.

The biggest problem in education is not, and never will be, that too few students have access to Harvard. Let’s ignore for a moment the discussion about the value of an Ivy League education. Instead, consider the myopia of this view of education. K-12, as opposed to higher ed, is more relevant to the vast majority of Americans. Rightfully, getting every kid into Harvard is far from the top concern of anyone involved in this space, whether as a student, parent, educator, policy-maker, or entrepreneur. Let’s get every kid into and through high school first.

The many issues in K-12 –the achievement gap, funding disparities, and teacher attrition rates, to name just a few –are not issues that can be resolved by instant access. In fact, they are not software problems. They are human issues, political issues. Software alone will not change ,  much less save ,  the world of education.

Software can help, under three conditions:

1. Edtech should be designed to focus on a narrow set of problems.The complexity of the American education system can be frustrating for companies, but it exists for a reason: educating 77 million students isn’t simple. Certain parts of it must exist to manage so many people. Companies aiming to overhaul education with their product are DOA. Instead, they should be targeting inefficiencies in the system. Companies ought to recognize that there are certain confines within which they must operate.
2. Related, edtech companies ought to make sure the problem they solve is a real problem, not a mere annoyance.Often, great edtech software never gets adopted because it’s solving a problem that doesn’t yet exist, or won’t exist until 20 years of innovation have transpired. Companies that believe that their software will help to usher in a whole new educational model, who will truly disrupt or “eat” education, probably won’t. Iterative –not radical – change is itself ambitious in education, an industry in which the most successful companies thus far reinforce the status quo.
3. To focus on the right problem, companies must assume a purely Socratic approach :  know that they “know nothing,” and be inquisitive, open-minded, and responsive . By asking rather than answering questions, companies can deliver software that helps humans improve teaching and learning.

Using good software, humans begin to chip away at the big issues of the day: software might free up time to focus on the issues, surface data that elucidates the issue, and empower teachers and learners to grow in new ways But humans do the important work. Software must strike a delicate balance between innovating and meeting current, non-technological needs–a balance that’s harder to strike in education than anywhere else. But who said startups should be easy?