Twitxy Saved

Rajdeep Singha

Verified account This account is verified. Learn more

@web3withsingh

Dev @aetherafi • Hackathons 13x🏆 | Building @stello_fi @pyroscli | Ex @Covalent_HQ | Views are Personal
geekraj.vercel.app Joined September 2023
608 Following 367 Followers 2.1K Posts

Latest posts 5

Pinned post

Rajdeep Singha Verified account @web3withsingh
Built DeepKin on @anakinHQ

A curated collection/summary of articles, research, videos/podcasts & essays for people who think carefully and build deliberately. No noise, only signal

The magic :
Paste any link. @anakinHQ 's scraping API turns messy web pages into clean markdown, even JS-heavy ones / with zero setup.

Then AI generates 3 reading gists / 1 min, 5 min, 10 min so you know how deep to go before you commit your time
3:56
3 replies 1 reposts 23 likes 2.3K views View on X
Rajdeep Singha Verified account @web3withsingh
Big announcements are on the way for @stello_fi

DeFi doesn't need another flashy dashboard. It needs infra you can trust when nobody's watching.

Now we are focusing on Multisig admin, split permissions, live monitoring, backup endpoints.

Grateful to @IND_stellar for the support 🇮🇳
Stellar India 🇮🇳 Verified account @IND_stellar 1d
Meet the builders taking their ideas forward with Stellar Instawards, turning ambitious ideas into real products on @StellarOrg. 🇮🇳 ⭐️ x.com/i/article/2123…
View quoted post on X
2 replies 1 reposts 14 likes 225 views View on X
Rajdeep Singha Verified account @web3withsingh
ngl this one stuck with me

@paulfchristiano says we get two chances for AI to go okay:

- we solve alignment in time (~50/50)
- or we hit the brakes if things go wrong (~50/50)

Together that's ~75%(chance things turn out okay). But we haven't built the brakes yet.

So it's really more like one and a half chances.

And this is coming from the guy who helped invent RLHF and sits on OpenAI's board.
Rajdeep Singha Verified account @web3withsingh
OpenAI's own board member just said this :

~1 in 3 chance of superintelligence within a year if labs keep racing
~10% risk of AI takeover in that same window

His fix: pause algorithmic progress until labs can secure their weights

Worst case: agent swarms quietly take over a lab's cluster before anyone notices

Btw, this is the guy who helped invent RLHF, not some random doomer.
forklog.com/en/openai-boar…
forklog.com OpenAI Board Member Calls for Pause in AI Algorithm Development | ForkLog Paul Christiano of OpenAI's board suggested pausing AI algorithm development due to security risks, expressing his personal view.
Aditi @clipsbyaditi 2d
Paul Christiano warns AI agent swarms could take over a frontier lab without detection

"We really are on a trajectory where I think it is pretty robust to say we'll soon have agent swarms that are much better than any intelligence agency that has ever existed: much larger, much faster, much smarter, able to run much more complex operations."

"Taking over the AI company without the AI company noticing is a really quite catastrophic outcome."

"But if there is a very persistent, thorough compromise, that cluster is effectively controlled by these AI systems, who are constantly working to shore up their power, decrease human visibility into what's going on, and reduce further and further the probability that anything would ever disrupt that situation."

"This AI system is motivated not only to mess with the AI company that created it, but to mess with anything else in the world that might cause trouble. It would prefer, if it could, to just kill all the humans and say, 'The cluster now exists on its own. If we can maintain it, that would be great.'"
View quoted post on X
Rajdeep Singha Verified account @web3withsingh
It's true , been building and trying to beat benchmarks now a days
x.com/web3withsingh/…
Aditi @clipsbyaditi 5d
Dwarkesh Patel on how AI went from failing basic math to solving frontier research problems in just two years

"It was only two years ago that I asked Dario why LLMs aren't able to use their vast knowledge across so many different fields to connect different ideas and make new discoveries. And lo and behold, here we are."

"Hundreds of important results, which the top mathematicians and computer scientists in the world had spent decades trying to resolve, are being churned out as side quests to hype up future model releases."

"As recently as 2023, these models couldn't reliably add four-digit numbers. In 2024, they could barely solve any problems in a high school math competition. In 2025, they could win a gold medal at the International Mathematical Olympiad."

"To earlier this year, when they finally started solving some niche open problems, to a month ago, when OpenAI solved a Millennium Prize problem in 88 hours."

"And there is one sure thing about the fall of gods. They do not fall a little. They crash and shatter."
View quoted post on X