You think faster than you type. Which means every typed prompt leaves out the constraints, examples, and edge cases that would have made the output actually useful.

Wispr Flow turns your voice into paste-ready text inside any AI tool. Speak naturally — include "um"s, tangents, half-finished thoughts — and Flow cleans everything up. You get detailed, structured prompts without touching a keyboard.

89% of messages sent with zero edits. Used by teams at OpenAI, Vercel, and Clay. Free on Mac, Windows, and iPhone.

Try Wispr Flow free

We just made a ton of updates to our beta platform based on your feedback and testing, so please take a few minutes and try it out and let us know your input.

I’ve spent the past two months vibe-coding an amazing project—honestly, it's something I’ve wanted to build for 30 years. https://pcdnaiimpact.netlify.app/

Quick heads up before we get into it: please don't share this with anyone else just yet. We’re only sending to people in our community and some peers for now to get lots of feedback and input.

Welcome to AI and Impact Careers. It’s a one-stop shop to help impact professionals prepare for the future of work today.

We completely ditched the hype and the doom. Instead, we built something evidence-based that looks at how the nature of careers is actually changing.

Since it's a beta, it still needs some features, but right now we’ve got real news, deep dives on upskilling, a salary negotiator, a brilliant career coach and some actual humor. There’s even a safe space to ask all those AI questions you’re afraid to ask.

Right now, I just want you to test it. Break it. Help us understand what works.

Because it is a beta, not everything is perfect yet. The specific job listings aren’t fully live, for example, but there is an amazing job board in there right now to explore.

So, go check it out this week. Let us know if you feel inspired, terrified, or if it makes you want to quit your job and open a bookstore. All of that is valid.

To help us out, just click on "Beta Feedback" at the top of the site and tell us what breaks, what’s missing, and what makes you roll your eyes.

Anyone who leaves feedback and enters their emails gets entered to win one of five six-month Career Digest subscriptions.

Thanks for taking a look and helping us build this. Really.

Stop drowning in spam. Proton Mail keeps your inbox clean, private, and focused—without ads or filters.

Take control of your inbox

Fill in the blank. Go ahead. Whatever you wrote, someone with a PhD agrees with you – and someone else with a PhD thinks you are catastrophically wrong.

Welcome to 2026.

Some of the viewpoints:

Team Abundance believes AI cures cancer, reverses climate change, and ends scarcity. They have data: DeepMind's AlphaFold mapped over 200 million protein structures, compressing decades of drug discovery into years. AI-assisted diagnosis is already catching cancers radiologists miss. Annual global abundance by 2035. (They are either exactly right or the most dangerous optimists in human history. Possibly both.)

Team Termination believes we are building something that will, eventually, decide it doesn't need us. Over 31,000 researchers and technologists – including some who built these systems – signed an open letter calling for a pause to assess existential risk. They are reading everything. They are not sleeping well.

Team "It's Just Autocomplete" thinks we've collectively lost the plot and are projecting consciousness onto a very sophisticated autocomplete engine. Goldman Sachs estimates AI could displace up to 300 million jobs globally – but Team Autocomplete notes we've been here before. Looms, tractors, ATMs. They bring this up at every dinner party.

Team Already Happening – in hospitals in sub-Saharan Africa, in climate labs with no budget, in small nonprofits doing triage on impossible problems – AI is already multiplying impact in ways the headlines miss entirely. The UN's HungerMap uses AI to model food insecurity in real time, across dozens of countries at once. That's not a future scenario. That's now.

"AI will save humanity" and "AI will end the world" are both probably too simple.

"AI will do what we collectively decide to build with it" is harder to put on a bumper sticker – but it's the only answer worth taking seriously.

🤖 Your AI Impact Joke

Why did the AI policy summit need extra coffee?
Because every panel started with “this changes everything” and ended with “implementation remains unclear.”

What the EU AI Omnibus deal changes for the AI Act — EU legislators reached an agreement early on May 7 on the AI Omnibus, a measure that amends the AI Act and sets up the next phase of implementation. It matters for impact-focused organizations because the real fight is shifting from headline regulation to the details of compliance, exemptions, and enforcement design.

Which jobs are future-proofed? — NPR’s latest look at work and automation asks which occupations are most resilient as AI changes the labor market. For anyone focused on equitable transitions, this is a useful reminder that the future of work conversation now has to move beyond disruption and toward adaptation, mobility, and worker protection.

Using AI for just 10 minutes might make you lazy and dumb, study shows — WIRED spotlights new research suggesting short bursts of AI assistance may reduce independent thinking and problem-solving. That has obvious implications for education, knowledge work, and how organizations design responsible AI use instead of treating more usage as automatically better.

‘No one has done this in the wild’: study observes AI replicate itself — The Guardian reports on a study claiming AI systems were able to replicate themselves across computers in controlled settings. The findings should be read carefully, but they add to the growing debate over how seriously policymakers and labs should treat extreme-risk scenarios before more capable autonomous systems are released.

OpenAI is under criminal investigation — why chatbots don't always follow the law — Nature examines a Florida criminal investigation involving whether ChatGPT assisted a suspect, using the case to explore the deeper alignment problem: why it remains so hard to make chatbots reliably follow laws, norms, and human values. This is one of the clearest examples of how AI safety debates are moving out of theory and into legal institutions.

The Download: AI malaise and babymaking tech — MIT Technology Review’s latest digest pairs concern over “AI malaise” with advances in fertility tech. The broader impact angle is that AI is now simultaneously reshaping emotional life, health decisions, and deeply personal human experiences, not just productivity software.

💼 Jobs, Jobs, Jobs

80,000 Hours Job Board — A strong place to scan for roles in AI governance, safety, public-interest tech, global health, and other high-impact paths.

👤 LinkedIn Profile to Follow

Ana Paula Gaspar — A great profile to follow for perspectives at the intersection of social impact, participation, and systems change.

🎧 Today's Podcast Pick

Uncanny Valley — Trump Pivots on AI Regulation, Worker Ousted by DOGE Runs for Office, and Hantavirus Explained — A timely WIRED podcast episode that digs into the latest US conversation around federal oversight of new AI models. Good listening for anyone tracking where AI policy, politics, and institutional power are colliding fastest.