AI Lets Me Game More - and Work Better

In this Web News episode, the hosts explore what people should do during the downtime created when AI agents are running tasks—which can take anywhere from 5 to 40+ minutes depending on the prompt complexity. Rather than advocating for maximum productivity during these gaps, they argue for strategic breaks and leisure activities as a means to improve overall output quality.

Mike introduces the concept by sharing his recent habit of playing Game Boy games (specifically Pokemon ROM hacks) on a handheld retro gaming device during AI agent runs. He explains that he discovered taking these 15-20 minute breaks a couple times daily actually helps him return to work in a more focused mental state, leading to better decision-making and quality output. He contrasts this with his previous behavior of attempting to run multiple agents simultaneously, which left him feeling exhausted and resulted in degraded work quality.

Matt expands on this theme by discussing how he uses the in-between time to tackle other work tasks, get coffee, or use handheld gaming devices, framing these activities as relieving work-related anxiety rather than wasting time. He notes that while there's a limit to multitasking capacity (cognitive bottleneck), strategic use of AI to run parallel tasks while he handles other work creates a sense of progress that alleviates stress.

Both hosts push back against corporate culture that treats every minute as billable productivity, arguing that happier employees produce better work—a principle they claim has been studied for decades. They acknowledge the tension between corporate pressure to maximize productivity metrics and the human need for breaks and mental recharging. They suggest that small-to-medium businesses often take a more results-oriented approach, caring primarily about whether work gets done by deadlines rather than micromanaging how employees spend each minute.

The hosts position this as potentially the beginning of AI actually fulfilling its promise to relieve human labor, though they acknowledge this may be aspirational thinking. They conclude by inviting listeners to share what they do during their AI agent run times, questioning whether more aggressive approaches to productivity are actually optimal.