---
id: "concept-so-so-technologies"
type: "concept"
source_url: "https://hbr.org/2024/12/how-to-create-value-systematically-with-gen-ai"
source_title: "How to Create Value Systematically with Gen AI"
source_timestamps: ["§ Individual Improvements"]
tags: ["economics", "productivity", "automation-trap"]
related: ["claim-individual-gains-insufficient", "concept-value-creation-pyramid", "entity-daron-acemoglu"]
definition: "Innovations that displace human labor but do not generate sufficient productivity gains to improve overall enterprise competitiveness or societal living standards."
enrichment_confidence: "medium-high"
sources: ["spine"]
sourceVaultSlug: "hbr-seg-spine"
originDay: 1
articleStem: "hbr-nm-98-create-value-systematically-genai"
sourceUrl: "https://hbr.org/2024/12/how-to-create-value-systematically-with-gen-ai"
sourceTitle: "How to Create Value Systematically with Gen AI"
---
# So-So Technologies

**So-So Technologies** is a concept coined by Nobel laureate economist [[entity-daron-acemoglu]] and his co-author Pascual Restrepo to describe innovations that *displace* workers but fail to increase productivity enough to meaningfully impact macroeconomic competitiveness or improve lives.

In the context of Generative AI, the authors — [[entity-todd-mclees]], [[entity-nicole-radziwill]], and [[entity-greg-satell]] — argue that isolated individual productivity gains often fall into this category when applied at the enterprise level. Cited examples of one-time task gains: a customer-service agent resolving issues **34% faster**, a software engineer writing **26% more code**, a data scientist completing tasks **10% faster**. While each is a significant improvement for a specific task, none fundamentally alters the organization's value proposition or operational model.

Organizations that stall at this stage merely make the technology available and *hope for the best* (see [[quote-hope-for-the-best]]), producing minimal aggregate effect and a high risk of lagging behind peers pursuing deeper transformation. Escaping this trap is the entire purpose of the [[concept-value-creation-pyramid]], and this concept underpins both [[claim-individual-gains-insufficient]] and the contrarian framing in [[contrarian-productivity-gains-are-insufficient]].

**Enrichment / validation.** Acemoglu & Restrepo's original phrasing is "so-so automation": automation that "does not bring about huge productivity gains but instead displaces workers and may reduce labor's share of income," and can drive wage stagnation despite technological progress. The extraction is a faithful paraphrase applied to GenAI. Their broader research contrasts this with **"task-creating"** technologies that augment labor and lift both productivity and wages — the lens under which Levels 2–4 of the pyramid are the "good" direction. A counter-perspective worth holding: for firms with large knowledge-worker cost bases, even a "so-so" 5–10% aggregate efficiency gain can materially expand margins, so Level 1 is better read as *necessary but not sufficient* rather than worthless.


## Related across articles
- [[concept-efficiency-ceiling]]
- [[claim-individual-productivity-roi]]
- [[claim-efficiency-not-advantage]]
- [[claim-individual-gains-insufficient]]


## Related across segments
- [[concept-efficiency-ceiling]]
- [[claim-efficiency-not-advantage]]
- [[contrarian-efficiency-trap]]