Working on what's next. Prev cofounder @AIUnderwriting and AIUC-1. Building positive sum. Via @Harvard @ThielFellowship

Joined August 2007
219 Photos and videos
I guess there really was a button they made for Tibo that eventually got good enough to just release…
Jun 12
We heard you wanted to use Codex rate limit resets on your own time. Starting today, we’re rolling out the ability to save rate limit resets to use later. We’re starting Go, Plus, Pro, and Business users with one free reset:
93
Brandon Wang retweeted
Fascinating and clever.
NEW: malware developers added nuclear & biological weapons text to to their spyware. Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner. Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky. When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit. We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted. In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation. H/T to colleagues that shared this with me socket.dev/blog/mini-shai-hu…
16
30
831
175,510
huge moment for my parents as it turns out that piping full sentences into Google with question marks actually yield better results now
1
71
if you get a whiff of codex about to reset limits, you can start 150 different breath.md sessions all at once to get some extra breathing in for free!
boris cherny goes on a podcast every three months and says something like “i’ve stopped breathing now i just wrote a breath.md” and the next day everyone in sf stops breathing
246
Brandon Wang retweeted
right monitor is 20 codex instances. left monitor has situational awareness on autoscroll. center monitor is my word doc mainfesto. two keyboards, one for both hands. left airpod is dwarkesh x eric jang, 3x speed. right airpod tchaikovsky. meta quest 3 overlays my HUD: heart rate, words per minute, blood caffeine content. one assistant hooks me to an iv of chinese peptides, cocktail. the other feeds me kimchi. my unitree robot steps in when my posture slouches. blue light beams down on me in my herman miller chair. efficiency. no wasted movement. no wasted thoughts. think you can keep up with me? good luck. this is just for my morning emails.
74
238
3,266
258,489
never have we been closer to the "everything app". my ChatGPT on iPhone is now an embarrassing collection of unrelated topics: - Codex: very important work!!!!!! - Images: random ghibli images - Library: no clue - Apps: ?? Recents - Stir fry recipe - Newark runways count
3
166
one of the things i've thought about since people I know have been dunking on the AI prompts of some prominent VCs ("you are an expert in everything") is that, probably at some point not that long ago, this was a reasonable thing to do-- old practices die hard
May 13
no bro you need to turn on “/extrausage”. dawg are you sure you have “/fast” mode on? Did you check the “no mistakes” toggle? are you sure you picked “correct mode”? did you turn up the “autonomy slider”, that’s how the pros use it,
1
222
Brandon Wang retweeted
The permanent underclass is coming… So we’ve helped 300K people train to be nurses, plane technicians, and AI-native at their jobs We have a new look and mission: to make people irreplaceable in an age of AI 👊
5
2
42
4,387
but i just read on twitter how people have been making a career out of brokering these deals! surely these upstanding individuals would never trick anyone right?
May 11
Anthropic just published a support page that should terrify anyone holding its shares on the secondary market. "Any sale or transfer of Anthropic stock, or any interest in Anthropic stock, that has not been approved by our Board of Directors is void and will not be recognized on our books and records." Void. Not restricted. Not pending review. Void. That means if you bought Anthropic shares through Forge, Hiive, or any other secondary platform without board approval, you are not a stockholder. You have no stockholder rights. Your transaction is invalid. It gets worse. Anthropic says it does not permit SPVs to hold its stock. Any transfer to an SPV is void. Investment funds claiming to offer indirect exposure are "most likely relying on mechanisms that attempt to circumvent our transfer restrictions." Forward contracts, tokenized securities, synthetic exposure products, all of it potentially worthless. Their advice to investors: "Assume that it is invalid." There is a multi-billion dollar secondary market in Anthropic shares right now. Platforms are pricing the stock at $265-$1,400 per share based on a $380 billion valuation. Real people have put real money into these positions. And Anthropic just told them none of it counts. This is the purest possible illustration of counterparty risk. You can buy a share of a company and have the company itself declare your ownership void because you bought it through the wrong channel.
5
457
Brandon Wang retweeted
May 11
me after 5 mimosas at mother’s day brunch yesterday explaining how I use claude at work to my grandma
21
269
6,128
425,583
That's a nice little password manager you have there. For saving that nice password we made you pick out when you made an account For sure you'll be able to use it as soon as you pick phone or email, these two buttons definitely do different things
1
105
Please @BrowserCompany or @OpenAI -- improve the Codex integration in Arc! The Codex extension loves to install into Arc but it doesn't work. Using both browsers is cumbersome.
1
89
Congrats on the launch! One really unfair thing is that the video conferencing platform knows exactly who is talking (and one of the reasons Gemini meeting notes is my current choice) — they just choose not to share I wonder if there is a vision (look for the highlighted participant?), browser DOM, or network based hack to pull this data out
Our #1 user request. You’ve wanted it for a while. I’ve wanted it too. Starting today, Notion AI Meeting Notes can determine who is talking in 1:1s and some video conferences! This is our first step forward and an active area of research, so let me tell you how it works.
1
292
AI resistant CAPTCHAs that require a human to scan a mobile phone!!! Oh no! cloud.google.com/blog/produc…
1
99
Jokes on you, I’m working on starting a fund that backs people improving the shovel selling experience
san francisco in one picture
3
289
Never has there been a more satisfyingly appropriate “create bucket” button
Ok but what if @FisherPrice designed @awscloud
3
185
Brandon Wang retweeted
When my wife asks how we are going to afford our trip
103
527
9,308
466,253
These guys keep on innovating-- super impressive! I really wanted to like @eightsleep and paid for a month rental (no free trial! I was bought in!) -- but I couldn't get the initial autopilot to work well. I'd wake up in the middle of the night cold :( Ended up returning
We have built an Autopilot feature to help women sleep better during a very critical time of their lives. Pregnancy mode was developed with research spanning 45,000 nights of sleep data from pregnant women. The largest research foundation of its kind. We're not stopping here. We're building for every stage of a woman's life. Rolling out May 10 in the @eightsleep app
1
294
what NYC people think dating in SF is like
2
124