Mic. Drop. 🎤 ⬇️

(edited to use a more neutral emoji)

https://youtu.be/h3AtWdeu_G0

This was sheer genius. Bernie got the chatbot to *completely throw these companies under the bus.* 😂

Now they're in a bind. Either they *agree* with the magical talking robot app, in which case they're politically screwed, OR they say the bot was hallucinating and providing false information & poor analysis, in which case they're undemining their wizardly thesis entirely along with their quest to usher in the age of AGI.

Bernie just completely pwned the tech bros!

@jaredwhite Sadly, tech bros will just shrug or say "lalalala, I can't hear."
@jaredwhite this is what would happen in a rational world. But this here is the lol nothing matters world, I'm afraid.
@hllizi @jaredwhite yeah, tech bros can just repeat phrases like "transform our lives" and "disrupt the world" and "shareholder value" a couple times, and everyone will forget what happened 5 minutes ago...
@jaredwhite as the other commenters - I fear for all the cleverness, people in general just won't care (or won't have enough time/energy to care), we're all kept so stressed and worried about so many things I just thinks human attention is spread too thin - and that is what all the bad actors are hoping for. There are too many fights to fight and everyone is too tired and too busy - the overwhelm is vast.
@jaredwhite this actually shows one of the other problems with AI, it's "yes and"-ing him, because that's what they do, adjust who they talk to and don't contradict unless prompted or fine tuned for it, you can get it to say almost anyhing that's not too outrageous this way (and just need some extra effort to get truly unhinged responses), so all this is only as meaningful as the person listening allows it to.

@jaredwhite my goodness. That is just chilling. That long pause right near the end is so…

What a brilliant video

@jaredwhite are you aware of the meaning of the inverted red triangle?
(your post was shared to my tl and I was about to instant block you, but I think that probably you just used a random symbol?)
@Schafstelze @jaredwhite I think they just used it to indicate the mic dropping (i.e., moving downwards)
@Schafstelze @jaredwhite
Um... it's the stock market devaluation symbol?
*shrug*
@roy_fireball @jaredwhite I would never claim that you play stupid.
@Schafstelze @jaredwhite
Good because I'm not playing. If it means something else you *can* just say it instead of dancing around the topic like everyone should just know it because spoiler alert, not knowing one piece of esoteric knowledge about some group co-opting a well known and daily use symbol does not mean one has a diminished capacity for thought.
@Schafstelze @jaredwhite
Oh, they blocked me... cool... I'll just stay ignorant forever I guess.

@Schafstelze I honestly haven't a clue…I did a quick Internet search and the only thing that I saw come up was some people who have used it in antifa movements since WWII?

I was just using it as an arrow pointing down. 🤷🏻‍♂️

@jaredwhite It's widely used by antisemitic accounts; originally the symbol is used by Hamas to mark 'targets' to be killed.
(it didn't make sense in the context you used it, so I decided to ask before an unjustified block)

https://en.wikipedia.org/wiki/Red_triangle_(Palestinian_symbol)

Red triangle (Palestinian symbol) - Wikipedia

@Schafstelze damn I had no idea. 😳 I'll edit my post!
@Schafstelze Your emphasis on the hateful use of a neutral symbol is good PR for the wrong group.
@gws what is the wrong group iyo?
(no answer needed, rhetorical question. Just go fuck yourself.)
@jaredwhite and we watch this on YouTube.
@jaredwhite Money, senator. The bluntness caught him so off guard it is shocking.
@jaredwhite That felt a little like Bernie had had many previous conversations with Claude and that, to a certain degree, Claude had his number and knew how to suck up. Even that great pause at the end seemed like a pivot only because it encountered pushback and needed to agree more with Bernie in order to keep him engaged.
@sstrader @jaredwhite Exactly. The people who are surprised or in awe about this video do not understand that these models say whatever keeps you engaged with them. It's the same as if some right-wing influencer posted a video of them asking grok about conspiracy theories and it agreed. Nothing of value was added, the models just repeat your own opinions and congratulate you on how smart you are. And I'm deeply in shock about how quickly "my bubble" now agrees with AI generated standpoints, just because they match their own biases

@gnarf @sstrader @jaredwhite I'm not surprised or in awe, but the video is rhetorically very impressive because it shows very clearly that Claude either (a) is accurate and contradicts the official narrative of Claude's capabilities and its societal impact, or (b) it's a sycophantic chat companion that can't be trusted for research or facts.

That's the genius of it. It doesn't even make an argument, it doesn't have to, it just exposes the absurdity at its core.

@jaredwhite This is an incredible video.

Senator Sanders is single handedly cornering the surveillance and AI industry politically.

@jaredwhite What would ChatGPT or Grok have answered?
@MichaelimOdenwald @jaredwhite Grok would have said that all companies are spying on their customers, but Elon does it best.
@jaredwhite So the LLM targeted at the liberal demographic can simulate liberal political talking points. That shouldn't come as a surprise. Bernie Sanders basically made an ad for Claude, where does a mic drop anywhere here?
@foyglgezang @jaredwhite after being very impressed I must say that this is the most relevant comment for me! The LLM just generated what the user wanted to hear. so imho- we must come back to think and reason for ourselves, senator Sanders
@jaredwhite
Damn, this was actually good and worth watching
@jaredwhite wow. Very impressed. I love Bernie's reaction when he gets the reply why the data is being collected. My reaction was the same.
@MaSch @jaredwhite but think of it: most people IMHO won't react like this - they don't see the threat to democracy or freedom. They do see the advantages, they'll tell you "Ive got nothing to hide" "my comfortable buying and selling experience is more important than the use of my data".
They do not see what's at stake here.
@energisch_ @jaredwhite yes, might be a point. but I doubt that MAGA people watch Bernie's Vids. In the end, the campaign produced CO2 and we know that LLMs can mimikry lefties.
@MaSch everything has downsides - and I personally will not use A.I. or A.I. agents: not for fun and I wouldn't have used them in my field of work before I retired.
But: these days you get A.I. assistance everywhere and you don't even know it. Till there is no control by law and watchdogs, we'll not know if we use it.
@jaredwhite

@jaredwhite

A privacy nightmare and unregulated... it is scandalous.

And a blind man can see it happen.

That is why the EU has the AI act btw… not our our watch!

@jaredwhite Imagine him being president. An actual guy with values.
@jaredwhite so easy to fake though

@dichotomiker @jaredwhite

Not for an entire career, no. Sanders was taking politically unpopular but righteous positions since day one. He has preached the same message regardless of what would be popular or potentially make him money.

Cynicism and skepticism have to be balanced by a willingness to actually examine and parse the facts in front of us. I don't agree with Sanders on everything, but there is no evidence that he's faking his sincerity.

@contrasocial @jaredwhite The dialog with an AI is easy to fake. The dialog with an AI!

@jaredwhite

It is sobering that you can read correct answers to his questions in countless blog posts and articles already.

After all, this is what this so-called "#AI" was trained on.

And still, people consider it "amazing" when a bullshitting machine generates a text from the knowledge it has ingested, possibly injecting inaccuracies and mistakes, so you better fact-check each of the sentences uttered here in this video.

The amazing thing is rather that people who could read what other, living human beings they trust have written … prefer to have a half-baked, half-true and minced version of it, read out loud to them by a smartphone.

@jaredwhite Great texter behind the Claude actor! 😁 We should teach people like this.
@NatureMC @jaredwhite
Initially I had the same thought - this could easily be faked. But that wouldn‘t match Bernie Sander‘s character. I would trust this conversation happened as shown.

@cpo64 I work with language and for me, the dialogue is too good. (I don't want to believe that Claude is so good and on the point but have not the time and money for a deep analysis).

I don't see that as a fake. It's an educational video, and if it's written by a human, it's a "literary" device. And the prompts are also forcing the way of answering. Intelligently asked!

@jaredwhite

@NatureMC @jaredwhite It is that good… this is the level of conversations one can have with LLMs these days.

@cpo64 Sorry, but I'm fact-checking texts if they are made with LLMs or by humans professionally.

It may be your feeling but it's not a fact. 😉

I’m not going to argue about who wrote the text either. As I said, I don’t have the resources to verify it beyond doubt, and I’m speaking here in terms of a probability based on experience. If a professional proves me wrong step by step (not by feelings), fine.

@jaredwhite

@NatureMC @jaredwhite That‘s interesting. I did indeed not notice any difference to other LLM conversations. Glad that experts can still identify differences.
@jaredwhite I've never had a conversation with Claude, does it really sound like Karoline Leavitt, or is that a preference selected by the user?
@jaredwhite @dgar from the Trojan horse’s mouth.

@jaredwhite the framing of this as an interview is so weird and anthropomorphizing.

You can get Claude to agree that you are a space faring cyborg wizard. Or that "Hitler had some good points". It should be no surprise that the engagement-bullshit-generator will iterate on Bernie's talking points.

@Spirit But I think that's the point! Bernie totally outplayed the bot's "logic" and got it to adopt his own talking points. So either he's right (I mean, yeah!), or the bot was very easily manipulated. Either way, it's a huge black eye against Anthropic et al!
@jaredwhite I believe the correct emoji sequence is:
🫳🏻
🎤
@marcintosh oh nice! I'll have to remember that

with Bernie Sanders doing the analysis.

https://youtu.be/h3AtWdeu_G0

Thanks to @jaredwhite for sharing this 🧵 🌹. I just added a CW as appreciated by my part of the Fediverse community.

Bernie vs. Claude

YouTube