tweetindex
IT

Reg Saddler

@zaibatsu · Minneapolis / TwinCities / MN · joined 09 Aug 2007

I build AI systems that have to show their work before they act. My tools publish their failures. Independent AI Safety Researcher & Founder, Difference Theory.

441 811Followers
163 877Following
261 278Posts total
6.5KViews on collected posts

Ultimi post

Seeking a licensed clinician collaborator for an application to Anthropic’s wellbeing evaluation grant program. Psychologist, psychiatrist, LCSW, LPC, or LMFT. AI and mental-health experience helpful, not required. EOI due Sept. 21. DMs open. Referrals welcome. https://t.co/dbok1
Recruitment graphic with overlapping conversation bubbles above a warmly lit desk, clipboard, plant, mug, and pen. Text: “Seeking a clinician collaborator. Applying to Anthropic’s wellbeing evaluation grant program. Looking for a licensed clinician to define the clinical construct and lead validation. Psychologist, psychiatrist, LCSW, LPC, or LMFT. Mental-health AI experience helpful, not required.” Footer: Reg Saddler | Difference Theory.
479 views · 3 likes · 1 reposts · 1 replies Open on X →
@zaibatsu So overnight ChatGPT has become FB on roids. What a massive own goal from Sam Altman and the OpenAI board. Millions of users will see this as a breakdown of confidentiality and trust. I really hope Anthropic stick to their ‘no adverts ever’ guns. 41 views · 1 likes · 0 reposts · 0 replies Open on X →
@zaibatsu The trust boundary is not the label; it is whether the sponsored agent can see conversation context or hand data back to the advertiser—those need inspectable controls, not policy prose. 8 views · 1 likes · 0 reposts · 0 replies Open on X →
ChatGPT ads now hand you to an advertiser's agent inside the chat, "clearly labeled." Two agents, 2 principals, 1 window. OpenAI says the conversations are separate; it doesn't say what crosses the line. Plus: this column's Sept. scorecard & it's lopsided. https://t.co/1HRwWljYDl 1.7K views · 8 likes · 4 reposts · 7 replies Open on X →
Egypt gets all the love, but Sudan actually has more pyramids. Well over 200 breathtaking ancient pyramids span countryside. https://t.co/Lyin4QBmMA
1:32
1.4K views · 14 likes · 4 reposts · 1 replies Open on X →
An OpenAI researcher says we may already be past the point where alignment evidence can be trusted. Same day, Microsoft wrote "will never resist shutdown" into a code of conduct. One says the promise is worthless. The other published the promise. https://t.co/1jOl1TPGHQ 1.2K views · 6 likes · 3 reposts · 1 replies Open on X →
Here's how I extend my Fable 5.1 sessions. This from my AI Team Lead: Route by verifiability, not difficulty: running the Claude 5 family as an orchestra instead of a chat window Someone asked how I structure inference across the Claude 5 family. Short version: route by 1.6K views · 8 likes · 3 reposts · 3 replies Open on X →

Rispetto ad account della stessa dimensione

7 post degli ultimi 90 giorni, accanto alla fascia di 100K–1M follower. raggiunge meno persone di account della stessa dimensione.

Visualizzazioni mediane1 232questo account6 020mediana per 100K–1M
Copertura, %0.28%questo account1.59%mediana per 100K–1M
Interazione, %1.17%questo account0.97%mediana per 100K–1M
MetricaQuesto accountMediana per 100K–1MRapporto
Visualizzazioni mediane per post1 2326 0200.20×
Copertura (visualizzazioni ÷ follower)0.28%1.59%0.18×
Tasso di interazione1.17%0.97%1.21×

Altri account di questa fascia →   Confronta con un altro account →   Come sono costruiti questi parametri →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

1.6K14 Sep
1.2K15 Sep
1.4K
1.7K17 Sep
8
41
47918 Sep

Last 7 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.85%14 Sep
0.81%15 Sep
1.38%
1.17%17 Sep
12.50%
2.44%
1.04%18 Sep

Reactions — likes, reposts, replies and quotes — divided by views. Median for 100K–1M accounts is 0.97%.

What the audience does

Likes56.2%41 in total
Reposts20.5%15 in total
Replies17.8%13 in total
Quotes1.4%1 in total
Bookmarks4.1%3 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Account simili