Weekly Link Out Loud (LOL) – W21 • 2025
Starting a new series of posts titled – Weekly Link Out Loud – LOLs – links to posts I read and took notes on from May 25 – May 31 2025
- AI-Fueled Spiritual Delusions Are Destroying Human Relationships: This is a great example of how the quest to understand ourselves with things like this is paved with nonsense. And it’s so easy and appealing to go down a wrong path Something to be mindful and skeptical about.
- Don’t Chop Up Your Apple Trees to Make Barrels for Your Apples: On TBC CEO’s notes that web pages won’t be the primary interface for a browser…
- How To Know What To Do: A guide on how to plan your life
- OpenAI has an unsubtle communications strategy.: The load-bearing relationships between futurity-vibes and finance
- Screen Names, BBM, and MP3s The Oral History of Millennial Tech: By the turn of the century, millennials found themselves as test dummies on the cutting edge of technology. Chat rooms! Music in your pocket! Tamagotchis! It all sounds pretty quaint now …
- The Anxiety-Curiosity Switch • Ness Labs: The anxiety-curiosity switch is the process of redirecting the mental energy that causes creative paralysis into curiosity-driven exploration. It’s particularly useful as anxiety and curiosity are reframes of very similar brain patterns. Doing this can often be the difference between creation and a slam of the metaphorical book.
- The Critical Difference Between Complex and Complicated: When you manage complex problems as if they’re complicated, you’re setting your company up to fail.
- Who You Are is Not What You Do: Rather than drastic, very difficult changes, get hobbies a little at a time, make friends a little at a time. Ultimately, it’s similar to diversifying a financial portfolio. You have to diversify your life. Diversify yourself
- Why I’m meditating 300 hours: This is an interesting follow. They’ve already achieved close to 100-150 hours and that’s might impressive. While I’ve mulled about meditation, I am continuing journaling.
#2025 #aggregation #blogs #curation #links #lol #posts #weekly #weeklyLol
Estimating the success of re-identifications in incomplete datasets using generative models [Nature][open access]
![A screen capture of text from a section of an online report. One section of the text has been highlighted by a viewer:
"...99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes"
The entire screen-captured text, including the highlighted section, is below:
"legitimate privacy concerns. Anonymizing datasets through de-identification and sampling before sharing them has been the main tool used to address those concerns. We here propose a generative copula-based method that can accurately estimate the likelihood of a specific person to be correctly re-identified, even in a heavily incomplete dataset. On 210 populations, our method obtains AUC scores for predicting individual uniqueness ranging from 0.84 to 0.97, with low false-discovery rate. Using our model, we find that [highlighted] 99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes.[/highlighted] Our results suggest that even heavily sampled anonymized datasets are unlikely to satisfy the modern standards for anonymization set forth by GDPR and seriously challenge the technical and legal adequacy of the de-identification release-and-forget model."](https://games.ngo/system/media_attachments/files/114/093/454/204/947/385/original/49b6ef5b9fd54e4f.png "A screen capture of text from a section of an online report. One section of the text has been highlighted by a viewer:
\"...99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes\"
The entire screen-captured text, including the highlighted section, is below:
\"legitimate privacy concerns. Anonymizing datasets through de-identification and sampling before sharing them has been the main tool used to address those concerns. We here propose a generative copula-based method that can accurately estimate the likelihood of a specific person to be correctly re-identified, even in a heavily incomplete dataset. On 210 populations, our method obtains AUC scores for predicting individual uniqueness ranging from 0.84 to 0.97, with low false-discovery rate. Using our model, we find that [highlighted] 99.98% of Americans would be correctly re-identified in any dataset using 15 demographic attributes.[/highlighted] Our results suggest that even heavily sampled anonymized datasets are unlikely to satisfy the modern standards for anonymization set forth by GDPR and seriously challenge the technical and legal adequacy of the de-identification release-and-forget model.\"")
Hey friends, my buddy Steve Thompson PhD (
