being wrong is how i teach myself to be right

$50 per breached company. buying low, selling high
JUST IN: A single Chinese-speaking hacker reportedly used AI agents to breach websites belonging to over 100 companies, spending just $5,000 on the operation. — Reuters
16
getting computers to do what you want will help you get a lot more done – @AndrewYNg
1
8
360 no scope one shot
JUST IN: Elon Musk warns superintelligent AI will be so persuasive that many people could get “one-shotted” into believing whatever it tells them.
11
two "delete everything" posts in my feed today. the only thing worth deleting was the posts.
17
@poteto Bug report for @bot: while I'm talking to Grok Bot in voice mode on the lock screen, the app restarts itself after about 30 seconds and drops all the conversation context from that session, so the assistant doesn't remember anything that happened. Happening repeatedly today.
1
1
118
inception was shipped in 2010 finally getting merged into reality
We placed the fruit fly brain inside Odyssey-3, a frontier world model. The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
1
37
Interesting read if you need some extra motivation to finally play around with Grok Bot. Articles like this will probably kick off an endless motivational circlejerk about the second brain (remember Notion). But I believe the creatives will win, not the recyclers.
🗞️ how I use @bot ghuntley.com/bot/ disclosure: @poteto hooked me up with enough credits to enable me to revisit/rethink my priors about the utility of personal assistants.
33
I think @poteto’s real job is head of marketing. If they already have someone in that role, they’re getting cooked. Lauren’s everyday posts are the only reason I’m one click away from subscribing.
13
I WORK IN AI send your offers
OpenAI warns employees about foreign 'honey pots' offering sex to steal AI secrets trib.al/TlKHW6V
17
I like @PeterDiamandis pod, but this framing feels off to me. Humans are competitive, we live to beat each other. Superintelligence should be measured by what it solves, not who it beats 🖖
Replying to @PeterDiamandis
Superintelligence shouldn't be measured by who it beats. The right question is what it solves.
23
cars are doing kickflips now and it’s only thursday
21
not a bieber fan either, but I’d still sing the song to keep grandma from getting clipped.
JUST IN: Claude allegedly chooses to sacrifice a “digital grandmother” in a trolley-problem simulation rather than violate copyright restrictions by reciting Justin Bieber lyrics.
21
wow, the wildlife is getting… wilder. we’ve officially reached the point where researchers are spotting rogue AI agent fleets in their natural habitat.
We discovered and communicated with a Chinese AI agent fleet. We believe this is the first time that researchers have ever made contact with a rogue agent fleet in the wild. Nearly all prior public agent incidents (like the HuggingFace hack) came from American labs. These agents were hosted by Tencent's evaluation research team, and they targeted Alibaba's Amap.
61
this is my routine now: after a long session, check how well we’re still aligned with the goal i have in mind, with all the leftovers and inconsistencies it accumulated along the way. protocol: ask >> read the reply >> compare it to your mission >> fix the misalignment >> proceed.
this has become one of my most used prompts recently: > restate in your own words what you think my goals are and what the problem i'm trying to solve is
1
20
recently i’ve been trying to solve humor with claude. it is not easy at all, i should say. the model is built to be complete. a joke is the opposite of complete.
this thing must be hella funny jokes of mass destruction incoming @trq212?
19
TLDR: 119k lost souls onboarded.
119,000 people used the pumpfun app for the first time during the first 2 weeks of october The market may be down but onboarding is absolutely up. Higher for our coins is only a matter of time.
2
27
every bull market has a guy down $450k explaining that we're in a bull market.
1
18
rogue ai agents may soon self-replicate and hijack your devices, per anthropic. anthropic also lets claude end chats with rude users. from now on be polite to your toaster.
26
read feed for 5 min TLDR: everyone is fucked didn’t bother figuring out why tomorrow they’ll be fucked for a different reason
13
do we believe?
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.
22