Ask what's the best use of your time and what an agent could do for you
Molly Graham suggests that when deciding what to work on, people should ask what the best use of their time is, and then what an agent or robot could do for them.
39,935 results · Most relevant
Molly Graham suggests that when deciding what to work on, people should ask what the best use of their time is, and then what an agent or robot could do for them.
Molly Graham argues we have to accept that, at least for the moment, AI models are a bunch of interns — a bunch of junior employees. She discourages outsourcing your vision or the future of the world to a bunch of summer interns, and recommends that you individually are phenomenal at a set of things — do not outsource that. You can enable it and amplify it, but don't outsource it to these weird robots that are effectively summer interns.
Molly's endorsed heuristic: AI needs the same coaching and training a human does — context, onboarding, many iterations — and behaves much more like a lazy intern than the best employee you've ever hired.
Molly Graham advises thinking not about what the form of your job was two years ago, but what it might be in five years — a concrete reframing question for career reinvention.
hedged
Molly Graham says one of the questions you need to ask yourself after typing something into existence is: is it good? — the slop-and-accountability question.
Molly Graham argues for breaking down the walls between traditional roles — designer, engineer, PM — that were important for the last 15-25 years. She says we need to shed a bunch of the old defined roles and ways things worked, because things are possible now that weren't possible before, and imagines a world with fewer walls between roles, making space for that inside companies.
Molly's endorsed stance for leaders: you cannot lead through the AI transition without talking about grief and normalizing emotions. Feeling a wild roller coaster of emotions during change doesn't make you crazy or broken — it means all the right things are happening. If you lead people without leaning into the creativity conversation versus the fear conversation, your team will be terrified.
Molly's endorsed stance: until there is real evidence jobs are going away, believe jobs are not going away but reinventing themselves — and take the entrepreneurial stance toward AI change, particularly in tech, until the data genuinely says otherwise.
Molly Graham says the data indicates AI will create a lot of opportunity rather than eliminate jobs, particularly for people willing to lean into reinvention. She suggests muting or ignoring the fear narrative around AI and jobs, and imagining your job will continue but become different — and that there may be more of them.
hedged
When a new technology launches, being one of the first to create listicles and blog articles about it results in ranking number one because there is no competition at that time. When AgentK.I launched, Ajelix was among the first to create listicles and blog articles for those terms and ranked number one. Launching in a market with zero competition before a trend goes mainstream means you can get into the first positions for those queries even with a super new domain and almost no backlinks.
Sophia Bush says that when announcing major life news like a marriage, you should take all the time you need to tell every single one of your favorite people face to face — fly to see them, show them photos, and be together in person before they find out online.
Atlassian built an 'AI Builders Week in a box' available on its website, showing the agendas they have used and how they utilized them.
Karri Saarinen says the goal for the future product organization is to keep the learning loop happening: have context, bring signals from that context, and bring those learnings into the whole team and company. He also hopes teams spend more time building understanding of what the company and product organization is doing, making sure the team is learning from something or someone — not just executing the next prototype or idea.
The firm's knowledge base is represented as files and handed to their experts to curate and update; it is a first-class feature in the product: a tab called Knowledge holds the full operational context the agents use, and because it is part of the product, people can edit the skill directly.
The firm is not doing autonomous trading with agents; agents propose trades and human experts decide whether the trade is right and execute it.
Contested by Lincoln Murr on AI Engineer
Give the AI agent file tools to read, search, and list files across the repository, so it can follow changes beyond the diff into pipeline definitions and access grants. Restrict the AI agent to repository tools only — it cannot query the live catalog, lineage tool, or production state; live facts stay layer one's job.
Configure the agent's rules by change type in the PR, knowing what layer one already covers, targeting query anti-patterns, sensitive data asks, and consumers outside lineage. For schema changes, have the agent walk the pipeline definitions and access grants in the repository to find consumers that were not registered in lineage.
When the downstream table needs to handle a wider type, update the downstream table first so its column can handle the wider type, and only then merge and deploy the PR. This ordering avoids incidents from type-widening changes.
When a column gets re-tagged as sensitive, search the code base for anyone still reading it from the copy that's about to lose it. This became a new rule after the re-tag miss.
Robinhood built a layered PR review system where deterministic scripts, an AI agent, and humans each handle different review responsibilities. About 60% of data PRs now merge without human review, but protected decisions still reach a person by design — the goal is less review work, not removing accountability.
Choose sources → to get a library of your own and a daily brief.