One of my favorite tools i've built for myself: a focus coach for my mac that talks back.
drift → it nudges, then yells
make excuse → it argues back
fried → it drops the roast and sends you on a break
meeting → it stays quiet
Focus app for Mac + Chrome, built for ADHD brains: an AI coach that catches you on YouTube, X or Reddit and talks (then yells) you back to work. Hit ⌘⇧Space and you can talk to your "live" calendar/assistant.
sooo good at keeping me on track. 👇
To recap: The FAA stopped every departing flight in the US today. It’s the first nationwide ground stop since January 2023.
By 6:40pm ET: 5000+ flights canceled, ~10,200 delayed. The FAA’s 5pm ET target to resume came and went.
The official cause so far: “communications issues.”
The AI part: 3 days earlier, the FAA expanded SMART to all 22 air traffic centers. It’s an $875M tool that suggests flight routes, and humans approve every suggestion.
In October, lawmakers relayed controllers’ warning that they had “little to no input” on it.
My guess: SMART didn’t do this. Software that suggests routes doesn’t usually knock out radios.
But that’s not the point. Hours into the first nationwide ground stop in nearly 3 years, the government can’t clear its own AI.
The routing layer, the part that decides who or what handles each job, is the most valuable piece of any AI system. It’s also the first thing everyone blames when something breaks.
If it can’t show what it decided and why within minutes, it shouldn’t be live.
Going from one region to 22 centers in one jump, after controllers said they’d had little to no input, is how you end up here.
The AI might be innocent. The rollout isn’t.
agi is not here...
as much as i love boardy, i love humans more
another alum i wanted to chat with gave him a 12-hour window and it showed me two slots. i said neither works. it sent two more (at 7 am for god's sake). we sorted it in one human email...
This is so terrible.
Beyond the personal controversy surrounding Jensen Huang, I find the broader implications of his departure equally unsettling. At a time when much of the frontier AI industry is increasingly concerned with restricting access to powerful models, controlling compute exports, and framing AI development as a geopolitical zero-sum game, Jensen has consistently advocated for open-weight models, global competition, and the democratization of intelligence. His vision of American technological leadership rests not on excluding competitors, but on building an ecosystem the rest of the world chooses to adopt. NVIDIA's proposed $12.93 billion acquisition of Hugging Face was perhaps the clearest expression of that conviction. His disagreements with figures like Dario Amodei extend far beyond open source, touching on AI safety, national security, export controls, and whether increasingly powerful compute should be treated as a strategic weapon.
What unsettles me is the larger philosophical divide this exposes, particularly at a moment when Anthropic is preparing for a potentially trillion-dollar IPO and the commercial stakes of controlling frontier intelligence have never been higher. Of course, Jensen has enormous financial incentives to advocate for openness, just as frontier labs have their own incentives to defend restrictions. I don't know whether these competing interests had any bearing on his departure, but I find it increasingly difficult to distinguish genuine concerns about public safety from corporate self-preservation and geopolitical ambition. Perhaps the most consequential question facing AI is no longer how powerful intelligence can become, but who gets to control it, who gets to benefit from it, and ultimately, who gets to decide its future.
NVIDIA letting Jensen go is a huge mistake, for America and for humanity. It's Apple firing Steve Jobs all over again, but much bigger, given the role of AI in defining the future of our world.
I've been keeping quiet on this out of respect for Jensen's family and team, who I've met and who are amazing humans, and because I personally hate infidelity. So my heart goes out to them in this difficult personal family moment.
But I can also say that Jensen is human, and makes mistakes like all of us do. Not that I matter at all, but I've known Jensen for a long time and have gotten the chance to spend a lot of time with him, behind the scenes, away from mics and cameras. I have seen him show kindness, humility, care to everyone around him, including strangers, including from the poorest most struggling parts of society, treating all the same with the same thoughtfulness and attention, because they are fellow humans and not because he has something to gain from them. That shows rare character, to maintain humility and humanity through all of it.
Another data point is what people who know him well have said privately to me about Jensen over the years. I've never had anyone tell me anything but how awesome he is or how much they love him. Probably over a thousand people at this point. This includes hundreds of engineers, CEOs, colleagues, friends, etc.
Also, since there was some spy talk. Jensen is as proud an American as it gets. I'm in shock frankly at all of this, including the mass hysteria around it. I love America, and I think NVIDIA letting Jensen go is terrible for America.
I stand with Jensen.
Love you all ❤️
I measured it with selective coverage: the share of tasks a model can decide alone while staying ≥95% accurate. Everything below its confidence threshold goes to a human.
Perplexity: 91.6%. Microsoft: 90.9%. Gemini: 70.5%.
Same accuracy as Jev, roughly 13 points less work it can safely take off your hands. It also costs more. Gemini runs $77 per 1M decisions vs $13 for Perplexity’s Decider, roughly 6× the price for a lower score.
Perplexity, Nace AI’s Drex Max and Microsoft scored higher than both frontier labs’ models, Google’s and OpenAI’s.
Google released Gemini Decision 1 this week, a model that scores a fixed set of options instead of generating text.
Benchmarked it head to head against six other decision models on 689 tasks. It placed 4th.
Every model scored between 89.6% and 93.3%, except Cloudflare’s at 79.8%. Accuracy barely separates them.
What does is calibration: does the model’s confidence match how often it’s actually right?
everyone’s arguing whether hydrogen beats argon. wrong fight. the number that matters: antigravity’s orchestrator switches carbon↔hydrogen mid-session and gets near-carbon success on 24% less quota, 19% faster. the model isn’t the product anymore. the router is.
Meet Gemini 4 Hydrogen: built for rapid coding, multimodal synthesis, and everyday workflows.
It finishes tasks 2.5x faster than 3.8 Flash and surpasses Argon on the @ArtificialAnlys Intelligence Index.
As someone who believes cheating is a fundamental moral wrong, and as a young Asian American who was drawn into tech by the extraordinary leaders who came before me, I can’t help but feel a certain tragedy in all of this, alongside an overwhelming fear for the industry. Perhaps America, in every era, has ultimately rewarded power more than hard work.
Today I killed the “Ignore limit” button.
Lock In, a focus coach that fights back:
👀 knows what you're actually doing
🗣 talks to you out loud (and listens)
👩 can sound like your mom, or your boss
💣 your friends can lock your laptop for 5 min
Launching soon. Mac + Chrome.