Mario ⌘

Recently, people have been telling me that they feel like they are too late to the AI game. I sympathize with this view, since there's real value in being early to a new technology, and incumbents have years of experience to build strong takes and credibility But I want to emphasize that it really is not too late. And there are talent bottlenecks everywhere. There is still room to scale existing orgs by quite a bit And short timelines are not set in stone. It's strikingly plausible that we will still have several years to prepare for the AGI transition, and it's all hands on deck rn I admit that it is pretty hard to get into the AI policy fellowships. They basically all have single-digit acceptance rates, which is pretty scuffed But if you do a bluedot course, join a SPAR project part-time, and generally keep up with recent developments in the field, I think you'd have a decent shot at getting into one of the fellowships. (so that's about 3 months of prep to get into a fellowship) And after getting into a fellowship, you have about 3-6 months until you'd likely get a full time offer. And after that, it maybe takes another 6 months to really get going. And after that, it's just policy impact baby Putting all of that together, if you start grinding today, you would probably have a full-time job in 6-9 months, and meaningful impact in 12-15 months And obv things could happen much faster than that

Major green flags in a non-linear candidate trying to break into applied AI: 1/ plays with coding agents every day 2/ tries to build before they know how it will all come together 3/ seeks out feedback from actual builders (engineers, designers, PMs) 4/ is curious and asks their agent to teach them things they don't know 5/ treats rejection like data and adapts 6/ figures out what goes wrong when their vibe coded creations fail 7/ actually uses git (they just tell their agent to do it when needed) 8/ tries out new models as they come out 9/ learns CS basicseven though they thought they were "non-technical" 10/ learns fast and stays focused on their goals 11/ reads their AI outputs and doesn't pass on slop to others 12/ thinks clearly, is precise with their words 13/ says "i don't get this, help me understand" without ego 14/ stays calm when everything breaks (it always will) 15/ keeps engaged month after month because they know this is the future any of these describe you or someone you know?

A few months ago Surya (@kickingkeys), a designer at Websim (@websim_ai) in New York, posted a project that went viral: “You can just RL a coding model to paint with javascript btw.” We sat down with Surya and were shocked to learn that he only wrote his first line of code in 2024 and had not heard of RL before January. We wanted to know how someone with that background goes from zero to training models in under two years, and whether anyone else could do that too. Our top 2 lessons from Surya: - Anyone can learn how these systems work if you are curious and willing to put in the time. - The best way to learn is to spend less time asking what AI can do to serve you and more time asking what it can do to serve the people around you. Some of Surya’s recent projects include training models to paint, setting up AI for a coffee shop owner so he has more time to write, and having agents watch him in the art studio and turn the footage into a reality-TV-style story. Surya is doing this for the love of the game. Read our full interview in the thread! cc @thisisrayouf

"We should be more ambitious with the software we build" This is a must-read post from @danluu for optimists looking to take advantage of this moment of AI innovation in software engineering TLDR: Software isn’t “solved.”. Better models just move us to more challenging work. What models can build vs what they can verify on their own will continue to grow. danluu.com/ambitious/

New essay: The End of Rationing sanketsaurav.com/writing/end…

Hi, I'm Felix. I'm an eng lead working on Claude. I love watching people take flight and do things they never thought possible. We often describe that as "being more ambitious" when using AI, which is tricky advice to follow. I wrote a tutorial: felixrieseberg.com/how-to-be…

the crazy thing about Cursor mobile is @rikcreation coded most of the real thing with Cursor, as a “designer” titles don’t mean shit. you can just build.