A round-up of recent “how is enterprise AI use and outcomes going?” links:
- Value seems to accrue to employee, the not enterprise. The ROI is, therefore, not legible. Related: Brian on “smart people” profiting from and using UI.
- The US military wants AI on every desk: “‘We want to have an AI capability on every desktop – 3 million desktops – in six or nine months,’ Michael said during a Politico event on Tuesday. ‘We want to have it focus on applications for corporate use cases like efficiency, like you would use in your own company … for intelligence and for warfighting.'”
- Bain not seeing huge gains in programming: “Two-thirds of software firms have rolled out GenAI tools, but developer adoption is low among those, and teams using AI assistants report a productivity boost of perhaps 10 to 15 percent.” Positive: $2,400/year ($200/month for Claude) is probably a low price to pay for 10% boosts in developer productivity.
- Murmurs of humans cleaning up slop-code, also at least one person in the SDT Slack.
- McKinsey is like “yeah, AI projects are all the usual shit to deal with."
- It’s certainly true that the more you bring to the AI, the better it will be. Also, if you bring some negative energy and wishes, you get that back.
- Outside of programming, the knowledge workers are having a rocky time too: “Employees reported spending an average of one hour and 56 minutes dealing with each instance of workslop. Based on participants’ estimates of time spent, as well as on their self-reported salary, we find that these workslop incidents carry an invisible tax of $186 per month. For an organization of 10,000 workers, given the estimated prevalence of workslop (41%), this yields over $9 million per year in lost productivity.”
- A UK government pilot reports some ambiguous results. From the original PDF: “The evaluation did not find evidence that time savings have led to improved productivity, and control group participants had not observed productivity improvements from colleagues taking part in the M365 Copilot pilot. However, many pilot participants reported noticing time savings in their own roles due to M365 Copilot.”
- “The future is already here – it’s just not very evenly distributed."
Three cases
Some theoretic takes in this thought-space:
- The positive take is that the workers now control the mans of ROI. The benefits of the AI are accruing to them individually. This is great for a worker! The problem is that individual productivity gains are not really legible (metric-able and report-able) to management. So, management cannot “see” the benefit. The other potential “problem” for management seeing the value is tha workers may not be doing more. They just can do the same amount of work in less time. So, the individual worker benefits greatly! They get paid the same and have to work less.
- The negative take is that there are not enough enterprise uses for AI.
- The middle take is that we just haven’t found them yet. As with the digital transformation era, management just needs to try harder, put programs in place, and change things. Might I recommend that you TryTanzi.ai?
I have positive feels for #3. It reminds me of Andrew’s joke about management wanting the benefits of DevOps: