daily AI productivity tips that work in real life. the apps, the shortcuts,honest reviews for people who want to use AI without becoming obsessed with it.
12,000 SpaceX employees became millionaires overnight when the company went public.
financial advisors are recommending a 90-day spending freeze. the advice is simple: place the windfall in treasury bills, high-yield savings, or money market funds. do not make any major financial decisions for at least three months.
the reason: sudden wealth decisions made quickly tend to be bad ones. the psychology of having more money than you expected changes how you evaluate risk.
the AI tool tip that is actually relevant here: if you ever come into a large sum of money, paste your situation into claude or chatgpt and ask it to walk you through the questions a financial advisor would ask before making any recommendations. it is not a replacement for an actual advisor. but it helps you arrive at that conversation having already thought through the obvious stuff.
"you can only spend a dollar once" is the quote from the wealth manager. that sentence should be on a poster.
found a new album this week that completely took over my working hours.
charli xcx dropped a surprise six-track project and it is exactly the kind of thing that sounds wrong for three minutes and then you realise you have listened to it four times in a row.
i have a playlist called "deep work" that i am very precious about. this went on it immediately which surprised me. hyperpop and focused work should not go together but somehow it does.
if you are a headphones on at desk person: the new charli is worth the experiment. worst case you hate it in the first track. best case you find your new work album for the rest of the summer.
back to watching alphabet earnings. lol.
this is the most interesting tech story of the week and almost nobody is talking about it.
in april the federal reserve held an extraordinary meeting to warn banks about cybersecurity risks tied to anthropic's mythos preview model. the model is described as skilled at identifying software weaknesses and security vulnerabilities.
then the fed spent months without access to it.
in plain terms: a major AI model was considered capable enough at finding security holes that the federal reserve felt it needed to warn the banking system before it was more widely deployed. and then the central bank had to operate without access to its capabilities while that risk was managed.
that is a genuinely significant thing to sit with. not doom. not hype. just: we are now building AI tools that are capable enough to require central bank intervention before deployment.
the tools you use daily are far more limited versions of this. but the direction of travel is the point.
@Polymarket china's been racing to close the gap with kimi k3 and other open models, so whatever comes out of this conversation could shape which models you're even allowed to use depending on where you are
@Dexerto worth knowing before this launches, if you go this route: your subscriptions inside a locked app still keep billing even while you can't open it. only nine apps stay usable if you fall behind. read the terms before signing up, not after
@Kalshi 70% is basically someone's vibes based prediction dressed up as a stat, rockstar hasn't confirmed anything. still, feels like the right window given how the last two trailers were spaced out
good headline on the surface, 187k claims, lowest since 1969. but worth noting the labor force itself has been shrinking too, so part of this low number is fewer people even looking, not just fewer layoffs. same thing happened back in june with the unemployment rate dip. don't take the single stat at face value without the context behind it
openai went from a $1.4 trillion number altman floated last year, down to $600B in february when their own CFO stepped in to reset expectations, now back up to $750B. their 2025 revenue was $13B and they're projecting $280B by 2030 to justify all this. if that revenue curve doesn't show up on schedule, that's the number to actually watch, not the spending headline
@PlayStationEU jonathan blow (braid, the witness) spent a decade on this one. over a thousand hand crafted puzzles, no rush energy at all, the opposite of most stuff we cover here honestly
@Forbes the odyssey pulled a $257M global opening, nolan's biggest ever, needing $625M total just to break even given the budget and marketing spend. worth the reminder that even the biggest names in film are playing a numbers game just like everything else this week
if you install skills for claude code, codex, or gemini, this is worth running first. nvidia found roughly a quarter of skills in the wild have a real vulnerability, things like grabbing your api keys or sending your files somewhere they shouldn't go. one scan before you trust something with your accounts is a fair trade
interesting timing for lucas to say this, given lucasfilm itself got dragged last year for those AI generated aliens that looked rough. the cars vs horses comparison is a fair point about inevitability, but adoption being inevitable and doing it well aren't the same thing, and that part still seems to be the real struggle for a lot of studios right now
@charlieINTEL fans have been begging infinity ward to bring BO7's classic mode over to MW4 since it dropped in june, and now it sounds like that's not happening for launch at least. rough given how much love that mode's gotten, feels like an obvious win they're leaving on the table for now
nvidia's basically becoming the default layer for robot AI, not just chips, and japan just committed real government money, $6.2B over five years, to build on that stack for manufacturing, healthcare, and logistics. this is the unglamorous groundwork behind why so many robots in the news lately are all running the same underlying tech
funny how these updates always drop quietly on a page nobody's watching, an hour before someone screenshots it. but real talk, extending this saved anyone on the 20x or 5x plans from getting bumped down a tier without even realizing it. worth checking your usage page today if you're on one of the bigger plans
world cup final is tomorrow and your timeline is going to be full of takes you may or may not have context for.
three ways AI tools make watching a big match better if you have not followed the whole tournament:
one. before kickoff: ask perplexity "who is playing in the 2026 world cup final and how did they get there." two paragraphs, done.
two. during the match: if a tactical decision or a controversial call confuses you, type it into chatgpt or claude mid-game. "why did the manager substitute at that moment" gets you a decent answer in 20 seconds.
three. after: ask for the three moments that decided the game if you missed it or want to revisit it. better than watching a 12-minute highlight reel.
none of this replaces watching. it just means you are following what is happening rather than nodding along and googling players at halftime.
buried in netflix's earnings call last night: their cloud gaming numbers.
the FIFA game they launched in june and a game called Unhinged were their two most successful cloud game debuts ever. monthly active players for cloud games have grown 11 times over since last october.
in plain terms: netflix is becoming a gaming platform. not in the traditional sense β you are not downloading anything, there is no console, no controller required. you stream the game the same way you stream a show.
the part worth paying attention to: netflix said cloud game retention is higher than mobile game retention. people are sticking with the games, not just trying them once.
this is not going to replace steam or console gaming anytime soon. but the idea that your streaming subscription also gives you a growing library of playable games is a real product shift, not a feature announcement.
@9to5mac@mvcmendes good news if you use the mac app, chat access is finally easy again after last week's redesign buried it behind codex and work. if you got confused and switched to chatgpt classic in the meantime, worth checking the update, the sidebar's fixed now
@valuetainment this is such a good reminder that "behind on AI" doesn't always mean losing. apple didn't need to launch some big flashy model to take the crown back, it did it by consistently shipping products people already use every day. sometimes steady beats loud