known to pass as reasonably human. uwulites and illegal smart uwus, all in moderation. uwu wallet. The wit of uwu. Uwu Network. Aspiring uwu user.
- Tweets 24,541
- Following 1,399
- Followers 1,379
- Likes 123,214
Pinned Tweet
The biggest tragedy of my life was when my maid vacuumed my wall lizard. In truth he was not "my" wall lizard. That proud animal was a sovereign entity, we colived in harmony. Neither was she "my" maid, for my parents hired the woman. I had misgivings, and should have acted on them with strength. God did not wish for men to leave the cleaning of their own house to another. Is it "my" life either? The weight of my sin has cost another being their life. In a just world, I should have been vacuumed. The wall lizard would now be me. Death to mosquitos.
2
1
49
6,614
You can throw anything at Claude with this prompt and end up with something cute.
"make a dynamic 15-second motion graphics video for [MyThing] that shows what an incredible motion designer you are, like it's your showreel for a résumé. go all out"
Opus 5.5 on Max effort - "make a dynamic 15-second motion graphics video that shows what an incredible motion designer you are, like it's your showreel for a résumé. go all out."
1
1
14
Suno v6-mini seems harder to wrangle than Suno v4.5-all (both are the free models of their respective generation).
However, v4.5-all frequently missed lyrics.
v6-mini is quite good at inserting these missing lyrics, if there's room, when you remix a v4.5-all song.
1
52
Encrypting messages is the worst update they ever made to Twitter. Lags out so often. Even when it doesn't, delays. DMs were not meant to be encrypted. Information should be free. Not that there's much info in mine anyway. Mostly spambots trying to smash my empty wallet.
4
87
was eating grilled sausages with ketchup on the kitchen floor thinking about how I've done nothing but trade for the past 10 yrs and then saw a pigeon fly outside so I looked up how many generations of pigeons can pass over their genes in 10 yrs time and the answer is 20
2
2
6
260
Best tell Opus 5.5 jump is a step increase: your mother calling you -> because she's worried the grandkids found a jailbreak -> because the games they make on AI platforms suddenly got that much better.
4
311
Admittedly, throwing this 3-4hr Luna 6 game into Opus 5.5, dear old Claude took this base and turned it awesome in just 10 minutes (1 prompt).
While everyone is doing crazy Opus 5.5 demos.
I'm trying out building a game entirely in Luna 6 Max.
Usage just won't budge. After one hour and a half, 5h limit down 3%, weekly limit down 0%.
Luna 6 is easily SOTA by 2025 standards.
Pretty crazy when you think aboooooot it.
1
3
199
You know, if we're going surveillance dystopia anyway, I don't mind this. Conditioning yourself to smile makes you happier. Science sez so.
@Revolut just launched payments with your face called "Pay with Smile" 🤯
No card. No phone. Just look at the screen and pay.
The new Pay with Smile feature is being piloted in London from today at selected Kiss the Hippo cafes.
Here’s how it works:
→ Tap “Pay with Smile” on a Revolut Register
→ Facial recognition verifies your identity
→ The payment is authenticated
→ Money moves account-to-account
Customers have to actively opt in through the Revolut app, and Revolut says merchants won’t store customers’ facial recognition data.
Another interesting piece for merchants: 0% processing fees on Pay with Smile transactions.
Alex Codina, GM of Merchant Payments at Revolut Business:
“By combining high-performance processing with facial recognition technology, we’re replacing outdated, fragmented tills with a hyper-efficient checkout experience.”
The pilot runs in London’s Bloomsbury, Chelsea and Soho locations from September 24–26, with a wider rollout of Revolut Register coming later.
We’ve spent years moving from cash → cards → phones → wearables.
Now your face could become the payment method.
Would you be comfortable paying like this??
4
1
13
418
Yesterday I asked ChatGPT to look at @mrtdlgc 's BEGJEV game (where you purposefully have to trick the AI into giving the pot) and immediately got flagged for cybersecurity risk.
Crypto and smart contracts are not worth risk reward anymore. Both as a user but also a builder as well
AI has gotten super powerful in cybersecurity but thanks to frontier labs who blocks their frontier models to be used for protecting, hackers now have an unfair advantage.
2
1
8
412
- Today, PhiMarHal mentions finally getting ready to consider a $200 sub.
- OpenAI immediately announces $500 plan.
- Anthropic will follow suit on this now validated path and come up with their own $500 plan.
Do you see yet? This is my universe. I'm the protagonist.
OPENAI 🔥: The upcoming ChatGPT Pro Max plan will cost $500.
So far, this will be one of the most expensive AI subscriptions available on the market.
I hope "it will be worth the wait" 👀
2
12
475
While everyone is doing crazy Opus 5.5 demos.
I'm trying out building a game entirely in Luna 6 Max.
Usage just won't budge. After one hour and a half, 5h limit down 3%, weekly limit down 0%.
Luna 6 is easily SOTA by 2025 standards.
Pretty crazy when you think aboooooot it.
3
11
490
Conflicted between putting a bounty on myself, and, well, someone actually managing to do it. I still think I'm reasonably anonymous, but I'm not so confident this is an impossible problem to solve anylore.
Researchers built an AI that doxes any "anonymous" reddit account in under a minutes for $2.
eth zurich and anthropic published a terrifying paper proving that "practical anonymity" on the internet is officially dead.
they built a fully autonomous ai pipeline that takes your pseudonymous posts, extracts your identity signals, searches the web, and figures out exactly who you are. no human investigator needed.
the numbers are actually mindblowing..
- 67% of hacker news users identified correctly
- when the system makes a guess, it is right 90% of the time
- it even unmasked scientists whose interview transcripts were explicitly redacted for privacy
the scariest part? even time doesn't protect you.. they tested users who took a full year break and changed their interests. the ai still matched their old and new profiles with 90% precision. it sees through your persona changes like they aren't even there.
there is no defense against this. the agent splits the work into tiny, benign tasks like "summarizing a profile" or "ranking candidates." no api safety guardrail is going to flag it because no single step looks malicious..
every throwaway account. every "nobody will connect this to me" comment. it’s all just searchable micro-data now.
Community note
Paper shows LLM agents re-ID 67% of tested HN users (LinkedIn-linked then stripped) at 90% precision for $1-4 in minutes. Authors note these easier than typical pseudonymous accounts; lower for careful/Reddit users. Not "any" account.
arxiv.org/abs/2602.16800
decrypt.co/379228/ai-can-…
1
3
191