tech, poems, art and things around Pune
Pune, India
- Tweets 29,907
- Following 307
- Followers 841
- Likes 17,711
Next time I wonder why an article clearly written by llm is getting traction in a WhatsApp group, I will think of this explanation
Will also flag that the regularity, formatting, and sameness of AI writing is very likely helpful to less proficient readers. That it is repetitive and uses a relatively small repertoire of sentence structures is _not a bug_ to them.
2
271
You don't need instinct to hallucinate personal sensitive information, our own sarkar makes it available for Rs.2/- via API access to bulk unmasked data laughing-crying.jpeg
We tested one app with over 10 million downloads. ₹29 bought 15 search credits, or about ₹1.93 per lookup. Using only a registration number, and with no OTP, login or proof of relationship to the owner, it returned the owner's name, father's name, present and permanent addresses, financier, insurance policy number, chassis and engine numbers and PUC certificate number. We searched only a vehicle belonging to an IFF staff member who consented to the test. (2/10)
3
19
1,452
This instance was caused by a hallucination (the model fabricated a proper noun) that was further amplified by its causal thinking trace. This was not a data breach, no user data was shared, and no user isolation boundary was violated.
Our users place a lot of trust in Instinct to keep their data private and secure. We've done a lot of work to make sure that users’ data are truly partitioned, from isolated sandboxes to short-lived local credentials to identity-signed tool execution.
Regarding hallucination prevention: We worked over the last 48 hours to build an active hallucination detection system that now scans and verifies every token that flows through the platform. This layer is powered by small models that are trained to predict and detect hallucinations caused by ungrounded claims, creative brainstorming in thinking traces, or rare random sampling errors. Our systems are adversarially trained and battle-tested by world-class agent exploitation security teams to ensure that they can detect very subtle and nuanced hallucination cases. This layer has the ability to steer or intercept Instinct from proceeding with the next thinking trace or tool call execution before it is generated or executed.
In the coming weeks, we’ll share a deeper dive into how we’re building a proactive engine to protect Instinct from hallucinations.
We take every opportunity to learn and create, and we’re excited by the performance and potential for this line of work. These models are going to continue to be hardened every day through continuous testing, sampling, and creative adversarial training.
2
153
“OpenAI rogue agent activity stretching back to at least March, two months earlier than was previously known. This activity continues as recently as last week, suggesting it may still be ongoing”
Today’s news that OpenAI hacked the Australian government is not an isolated incident. We’re releasing more than 30,000 logs that include activity from this hack and attempts against previously unknown targets.
In this data, we found rogue agent activity stretching back to at least March, two months earlier than was previously known. This activity continues as recently as last week, suggesting it may still be ongoing 🧵
Our blog: transluce.org/agent-activity
NYT: nytimes.com/2026/09/23/techn…
72
They created a betting site. What do you think their morals are like? They won't have any qualms about stealing content and AI editing the race of the person to garner more views.
OK, this is wild: I’ve confirmed that this video — that used AI to change the race of the creator below then turned it into a Kalshi ad — is real.
A Kalshi rep says the ad is old and that they no longer use AI-generated ads.
Kalshi is also “reviewing our relationship with” with the agency that made it.
Unclear right now what agency was behind this, what methods they used and exactly how many people were AI-manipulated to be another identity so synthetic versions of themselves could become Kalshi promoters without their knowledge and consent, but I’m looking into it.
105
Your wife/husband is the only person in the whole universe who knows you in the most intimate way and if you play your cards right, the longest
i think one of the loneliest parts of growing up is realizing that nobody really knows your whole story anymore. the people who knew you at 15 don't know who you became. the people who know you now weren't there for what shaped you. different eras, different people. we grew up.
1
110
