MangoChunks365

Person | Reader | Thumbprint

[MangoChunks365 @MangoChunks365]

@SellTheTeamFJF @Ostrov_A “Before and after” what, exactly? October 7th.

[MangoChunks365 @MangoChunks365]

@InternetH0F Angry Birds IRL

[MangoChunks365 @MangoChunks365]

@MatthewUSAF @allenanalysis He’s just the socialist Mr. Beast. The rhythmic hypnotic head bobs and hand motions and everything.

[MangoChunks365 @MangoChunks365]

@cinefiley @MLJHaynes @PerMariam A trip should not have a logo.

[MangoChunks365 @MangoChunks365]

@NextUpHalperin @MarkHalperin The delivery sounded like you reading something you didn’t yourself wholly write. And the chatGPT tropes were distracting and weakened good insights. Maybe I’m totally mistaken, and again I love your good word. Honestly don’t know how you do it all!

[MangoChunks365 @MangoChunks365]

@NextUpHalperin @MarkHalperin I watch your content religiously, @MarkHalperin; it’s truly invaluable. But large parts of the monologue seemed chatGPT aided. Your ideas/outline/prompts, but then the algorithmic alliterations and “not …, but…” patterns. Am I mistaken?

[TONY™ @TONYxTWO]

This is powerful 🔥

“It has been one month to the day since Charlie Kirk was assassinated. I want to explain how that’s affected my life..”

“You can find [old] videos of me saying with my full chest, ‘f*ck God’, because I was so entrenched in my atheism..”

“Since Charlie Kirk’s death I have personally organized two candlelight vigils, I’ve attended 3 candlelight vigils, and I’ve been to church every single Sunday since..”

“I’m gonna make sure that my son grows up admiring men like Charlie Kirk. And I’m gonna make sure he grows up believing in Christ.”

[MangoChunks365 @MangoChunks365]

@Skyislandia @islandgirl3961 @lenalovegood831 @Pontifex That’s a screenshot from a video of a police officer dad playfully arresting his child.

[🇺🇸 ʟᴇғᴛ ᴄᴏᴀˢᴛ ᴠᴀɢʀᴀɴᴛ 🇺🇸 @Baklava_USA]

Breaking it down for Democrats like they are 10 years old

This is why this black baby was zip tied you lying LEFTISTS

[Luke Person @LPersonTX]

@CTULocal1 “We can’t mourn Charlie Kirk he was hateful.”

“We’ll gladly mourn and eulogize a murderous cop-killing terrorist.”

These two things are being said by the exact same people.

[Sana Ebrahimi Ledene @SanaELedene]

I listened to Erika Kirk’s full speech at the memorial, and I want to share a few thoughts that came to me while live streaming the event. This is not political.

First, I should say that I grew up as a Muslim in a Muslim country. I don’t know enough about Christianity to say if what I witnessed is rooted in faith or culture. But what struck me most was how, even though death is heavy and this was by nature a sad occasion, the entire event carried a celebratory spirit that honored life.

That contrast hit me deeply. In Islam, even though we believe that good people go to heaven, the relationship with God is taught through fear. Funerals are overwhelmingly sad, often filled with warnings of the terrifying first night in the grave. Growing up hearing that, and then witnessing people celebrate life, speak of God’s love, and remember someone through the impact he had on others; it felt so refreshing, so positive.

Second, I was profoundly moved by @MrsErikaKirk’s words. I cannot fathom the strength it takes to stand and deliver such a meaningful speech after losing the love of your life. But even more than that, the grace it takes to forgive the very person who destroyed your world. I cannot imagine myself standing on a stage, sending love to those who cheered your husband’s murder, or inviting others to spread God’s love in response because, as she said, “we do not respond to hate with hate.” That is powerful beyond words.

Again, I am ignorant when it comes to Christianity, but if this is what it truly embodies, then I am envious of those who get to experience that feeling.

[Chris Ripa @CHRISsW0RLD]

They proved him right…

[Vatican News @VaticanNews]

📽️ *Leo from Chicago* 📽️
Vatican media present the documentary on the U.S. roots of Pope Leo XIV
A journey through the United States 🇺🇸 to explore, through voices, images, and testimonies, the life and figure of Robert Francis Prevost #PopeLeoXIV

Watch and share the trailer

[Monk Mindset @monkmindsets]

Harvard and Cambridge Neuroscientist, Sofia Carozza PhD, Gives Scientific Evidence Backing Monastic Life

[MangoChunks365 @MangoChunks365]

@rawsalerts “. . . in response to Iran’s nuclear attack” ? ? ?

[MangoChunks365 @MangoChunks365]

@stevensongs What could possibly go wrong?

[MangoChunks365 @MangoChunks365]

@gloria_purvis @librarythingtim So, does a good “why” sometimes excuse a bad “how”?

[Gray Connolly @GrayConnolly]

This may help many confused here...

[MangoChunks365 @MangoChunks365]

@SohrabAhmari Reminds me of, I think, R Brague expressing that a tradition is most itself when unaware of itself as tradition. Once something is consciously “traditional,” it has already changed profoundly.

[Luiza Jarovsky, PhD @LuizaJarovsky]

🚨 The video below is going viral, and it shows the result of running the following prompt 74 times on ChatGPT: "Create the exact replica of this image, don't change a thing."

Here's what it teaches us about LLMs, data protection & fax machines:

First, from a data protection perspective, we observe an accuracy problem.

A reminder that Article 5.1.d of the GDPR states:

"Personal data shall be: ... accurate and, where necessary, kept up to date; every reasonable step must be taken to ensure that personal data that are inaccurate, having regard to the purposes for which they are processed, are erased or rectified without delay (‘accuracy’)"

As I've discussed multiple times in my newsletter, and as we can observe from multiple examples of AI hallucinations, LLMs are not knowledge repositories.

They're merely predicting the next word in a sentence, based on the vast datasets on which they were trained.

The clip below shows another manifestation of the LLMs' lack of accuracy: the user prompt asked for the exact same image, but after every prompt (i.e., personal data processing), it changed the original image.

-

This user experiment also reminded me of fax machines.

Unexpectedly, fax machines and Generative AI systems might have more in common than most people think. I explain:

If you print a fax copy (I'm not even sure this is the right verb to use, as it's been 2 decades since I've used one) of a text and then reinsert it into the fax machine, every new copy will be lighter and lighter until it there will be nothing left.

This is what we're doing with the internet: Generative AI is being integrated into every existing system, and users are being pushed to use it for every single task.

The natural consequence is that the internet is being flooded with AI-generated content.

But there is a problem: LLMs are trained using data scraped from the internet. So if the trend continues, future AI models will be mostly trained with AI-generated data.

As recent studies have shown Generative AI trained with Generative AI trained with Generative AI leads to model collapse (similarly to the fax machine's blank page).

-

The user experiment also reminded me of the telephone game.

If you don't remember it, the game involves a child thinking of a word and saying it quietly to another child, who says it to another child, and so on. When the last child says the word, usually it has nothing to do with the original word (and everybody laughs).

Is this what we're doing to our information ecosystems?

I'm not sure everybody will laugh at the end.

-

👉 If you're interested in AI, especially how we make sure it's properly governed, I invite you to subscribe to my newsletter (link below) and join our 59,400+ strong community.

👉 The future of AI doesn't exist yet: we are creating it right now, and everyone should have their voice heard.

[MangoChunks365 @MangoChunks365]

@StaceyJSpiehler @Adi13 @rulajebreal Lawyers, paid actors . . . Hard to tell the difference in many cases, no?