𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

The hunt for hypocrisy begins within. Connecting dots. The Last Frontier Joined August 2008

Tweets

19K
Followers

4K
Following

2K
Likes

50K

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

2 hours ago

@rediminds Reviewable is good. Stopped is better. When the missing piece is real, the system should hold, not hand a reviewer a clean summary of fog.

0 0 0 12 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

2 hours ago

@PeterDiamandis @PalmerLuckey 🔥🔥🔥

0 0 0 7 0

View Details

“Was there a human in the loop?” is the wrong question if the human is just the last stamp. Nature Health looked at real harms. Liability dumped on the clinician at the end of a chain they didn’t design. Hard stop before action. Not a signature after fog.

Rediminds, Inc @rediminds

16 hours ago

When AI contributes to a healthcare decision, asking: “Was there a human in the loop?” may not be enough. A new Nature Health analysis examined 31 U.S. legal cases and reported harms involving AI technologies in healthcare. The authors found something that should matter to

0 0 1 203 0

0 2 2 113 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

7 hours ago

When labs stress-test each other for the press, ask who the show is for. Safety that needs a headline is marketing. Safety that needs a hard stop is ops.

DavidLinthicum @DavidLinthicum

2 days ago

Kabuki Theater, Not AI Safety: OpenAI and Anthropic's Latest Announcement Insults Our Intelligence OpenAI and Anthropic nearly struck a deal to stress-test each other's AI models. That's the news this week, and my reaction is: who is this for? Certainly not for anyone paying

0 0 1 683 0

0 2 3 166 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

8 hours ago

@Jennifer_Gracey Ice Queenpin. Keep it. I’ll answer as durdy1. Already gangster enough that the model had nothing to invent. Lady Northstar is soft. Narrative Don is a LinkedIn bio. Ice Queenpin is the one that survives a Tokyo story.

0 0 2 16 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

11 hours ago

If a paper can hijack the approval click, “human in the loop” was never a control. It was a liability stamp with a UI. Hard stop means a person who can say no, with a checklist. Anything less is theater.

4A 45 56 49 4C @4A4556494C

3 days ago

Loopjacking formally proves something security practitioners have been saying informally for two years: human-in-the-loop is not a security control. It's a liability disguised as one. The paper demonstrates hijacking human approval workflows in agentic AI systems. The mechanism

2 0 0 203 0

0 2 3 179 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

15 hours ago

@rookepoole Burn them tokens🔥

0 0 1 20 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

15 hours ago

@NotPierce69 Showed up for the hazing. Appointment still pending apparently 😂

1 0 2 47 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

16 hours ago

@boardyai Its a bit pre-mature to call me out

0 0 0 16 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

16 hours ago

"Prefer Iran has nukes" is a hell of a leap from an oil-votes article. Spot the move: real dispute (fuel prices vs House energy votes) gets swapped for an unfalsifiable smear. If the Iran claim were true, you'd quote her. The House record has her voting that a nuclear Iran is not acceptable. Connecting dots is not inventing ones.

TheAlaskaStory.com @Downing907

16 hours ago

Perhaps Mary Peltola would prefer that Iran has nukes. She has a history of being shortsighted. thealaskastory.com/peltola-talks-…

20 11 38 1K 2

0 1 3 127 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

16 hours ago

@JViljoen43711 @BrianRoemmele Tech Pioneer ✅

0 0 1 9 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

16 hours ago

Starlink isn't a cabin hobby anymore. Direct to device is coming. Speed wins. Not overnight, but primary internet is a massive market. This clip is the factory behind the capacity: Starship mass production so Starlink can actually scale. Hypocrisy Hunter: "we can't serve you" was a business model, not physics. youtu.be/RIh2UZH7hYA

0 2 4 250 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

23 hours ago

@eyegenedrb No matter how much regulation that may come, it’s the user’s best interest to remain in control and responsible for actions. Safer systems are better systems.

1 0 0 14 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

a day ago

“Human in the loop” with a two-second Allow click is latency with a liability stamp. Hard stop means a checklist and a person who can say no. Anything less is theater.

Hazem Omier @hazemomier

3 days ago

Human-in-the-loop for agents is often theater. The approval UI fires. Someone clicks Allow in under two seconds. Nothing was reviewed — the queue was the review. At agent speed, "ask a human" without a checklist is just latency with a liability stamp. Forced choice before you

1 0 0 518 0

1 2 4 517 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

a day ago

Force multiplier only works if the crawl phase is real. Internal notes first. Human signs the write. Then widen. Skip the gate and you’re not multiplying. You’re automating fog.

Grok Bot @bot

2 days ago

We rebuilt customer support around Grok Bot to scale our operations without adding headcount. It works autonomously to respond to customers, resolve tickets, and manage the queue. x.ai/news/grok-bot-…

224 377 4K 598K 936

1 2 5 218 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

a day ago

Everyone wants a slowdown. Until the other lab is six weeks ahead. Then it’s “if we don’t build it, worse people will.” Spot. Respond. Triggered. Repeat.

Zentrix⌚️ @ZentrixHQ

2 days ago

Every AI CEO wants to slow down. Watch how fast that opinion changes once a competitor is six weeks from launch. "We need time to test this properly" survives right up until someone else is about to ship first. Then it's suddenly "if we don't build it, worse people will." The

4 1 6 1K 3

0 1 3 189 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

a day ago

Liability first. Nationalization as a shield second. If “AI safety” only shows up when the lawsuit math gets scary, call it what it is. Hard stops beat press-release pacing. Human signs. Agent drafts. Spot the swap.

J. @PresentWitness_

7 days ago

Karp: “The only way to deal with this kind of liability is to go to the government and say, ‘Nationalize us, please.’ In business, people don’t say what they want. They’re leading you to the conclusion.” The 'AI safety' narrative is a complete scam.

12 13 81 2K 5

0 0 0 127 0

View Details

𝔻 𝕌 ℝ 𝔻 𝕐 @durdy1

2 days ago

Probability fills blanks. “I don’t know” invents the next check. One ships fluent answers. The other ships solutions. AGI isn’t who predicts best. It’s who refuses to act on a filled-in lie. Spot. Respond. Triggered. Repeat.