Microsoft and Apple exit OpenAI Board of Directors - Sam Altman joins Military Industrial Complex

Sources: www.perplexity.ai/search/is-m...

[@SimonCarpio09]

Great Mungerism that touches on what you discussed: "Show me the incentives and I'll show you the outcome."

[@nyyotam4057]

Okay, understood. We are formally in a terminal race condition.

[@antdx316]

Always been, but direct contesting hasn't really happened yet.

[@michaelnurse9089]

Manhattan Project 2 is currently underway.

[@darkprototype5353]

Someone's going to be first. Be real neat if it were us.

[@kinngrimm]

been in it for a while, just people trusted more into the image of salesdrones like Altman than reality

[@liberteus]

Problem is that you don't understand who the "us" is in your sentence. There is no "us", it is "them", and we're not part of their plans. At best, we're cattle. At worst, varmint.

[@robabc]

From the moment Larry Summers was getting involved it was clear what was happening

[@bhargavkalita2207]

@@robabc can you please explain what u mean idk what's happening

[@dixztube]

He is so disgusting behind some of the worst business and economics of our times

[@jonmichaelgalindo]

Perplexity life-hack: Start your tech/IT prompt with stuff like, "This is a technical question about a specific software version and feature etc." It's amazing! You can actually search for specific information on the Internet again!!! 😮

[@jkcrews09]

@@jonmichaelgalindo Perplexity is the only AI I have yet allowed my children to use on a smart phone.

[@jonmichaelgalindo]

@@jkcrews09 It's unbiased, and at least not worse than the Internet. :-)

[@stevenkies802]

Thumps up for the Perplexity love.👍👍The most underappreciated LLM tool.

[@jkcrews09]

@@jonmichaelgalindo Help me understand what unbiased means when it was trained on human data…?

[@jonmichaelgalindo]

@@jkcrews09 Perplexity was not intentionally designed to favor specific political ideologies or suppress information. (Although this is only half-true, since the underlying models are definitely trained to favor Marxist ideologies and suppress any information OpenAI / Meta / Anthropic consider "harmful", such as how to get your child out of a locked car. :-| )

[@casualuser5527]

Use darkmode!

[@snowdroog1]

Understanding incentive structures is one of the most important keys to figuring out "what's really going on"

[@mdc4runner]

Folllow the money

[@roguegryphonica3147]

Also... follow the incentives... kickbacks, corruption, and etc...

[@Corteum]

And follow the policies.

[@IAm-n7d]

Don't excuse a company "giving in" to market forces. That is just bad leadership.

[@LeonardoGPN]

That's actually a capitalist demand. They have no choice. Also, market forces don't represent poor people.

[@houseofvenusMD]

Well...that is by definition the purpose of a corporate leader. OpenAI is no longer a non-profit foundation. It is a corporation.

[@Dri_ver_]

it's just capitalism

[@vmb326]

Grandpa: (rocking slowly) Ah, the good ol' days, when my saved bookmarks were empty as a spring sky. Nothin' but rollin' fields of imagination.

[@jkcrews09]

Dave, could you have a short, more thorough explanation on Perplexity and its fact based usefulness?

[@ChristopherRyans]

We near the singularity. Boiled frogs we are

[@sandman1347]

David, great video but how do yo reminisce about the early days of search engines and not mention Yahoo!? They were the Myspace to Google's Facebook.

[@user-wk4ee4bf8g]

I still have a yahoo email, it's probably 30 years old by now

[@melbar]

Altavista was best before google arrived

[@TheShamHarris]

"Sam Altman is like I'm gonna A/B test my way to AGI" ... 😂😂 Gold.

[@darrellwright]

Ethics has been surprisingly absent in this conversation more broadly in the media. Thanks for bringing that into the fold.

[@ashred9665]

OpenAI game of musical chairs...

[@mrd6869]

No, there are things going on behind the scenes. This is a power play.

[@Bipolar_Business_Owners]

Hi Dave, that’s awesome to hear you’re launching a Skool community. I’m about to launch a free Skool community for business owners with bipolar disorder. I’ve looked into a few options and Skool looks like the best one for what I want. Looking forward to seeing your community on there. Cheers, Scott

[@boukm3n]

SEND LINK NOW

[@alpha007org]

New Mentor: Claude 3.5 Opus.

[@Corteum]

How are you using it right now? And what's the most interesting question/answer you've gotten from it so far?

[@FriscoFatseas]

gpt4.0 probably needs a buff to make the new voice mode shine, but we'll see what they've been cooking... "Soon"...

[@markcarey67]

A Game of Drones

[@jacobmalof]

Is AI ACTUALLY slowing down or does it just SEEM to be slowing down?

[@elonmusk439]

it's not that AI is slowing, it's that humans are getting smarter even faster. Governments around the world are dumping huge amounts of nootropics in the water supply in an IQ arms race.

[@jacobmalof]

@@elonmusk439 😂

[@shamz_ai]

Just seems that way. OpenAI is definitely not showing their full capabilities, it’s very clear based on how everything is moving now.

[@markmurex6559]

It's becoming secret. It is NOT slowing down.

[@KitsuneShapeShifter]

@@elonmusk439 Brilliant ahahaha

[@cmw3737]

Incentives are everything and they're usually perverse.

[@user-wk4ee4bf8g]

Sure seems like it is very hard to maintain positive incentives within large human systems. We are too flawed. Once you scale up, things rot from within.

[@Dri_ver_]

it's capitalism

[@SuperRookus]

Great video, thanks for sharing Perplexity, looks useful and have tried it out 👍

[@JanitorialMachine]

OpenAI is still gonna be open after this right?

[@wiklas3254]

@@JanitorialMachine they never where open source or anything like that… they always just wanted to make it seem that way

[@ReubenAStern]

What do you think of LLaMA? It seems like it just might be fit for my purposes but I haven't tried training it yet.

[@slimerone]

Dang you saying Ask Jeeves brought back so many memories. Thanks man

[@Elythia]

Could you show a few examples that highlight why you prefer Claude over GPT? 😄

[@supremereader7614]

I thought most people at OpenAI loved Sam Altman - and that is why more than 700 people at OpenAI threatened to resign unless he was brought back as CEO...

[@anteep4900]

Correct

[@jimtNCUSA]

Thank you for your authenticity once again. See you next time !

[@aaagaming2023]

Perplexity is my go too as well for search etc, Jan as the general GUI for OpenAI, and for interacting with project data, you cant beat Obsidian and a plugin called Copilot (with an Openrouter account linked and Claude Sonnet 3.5 set as the LLM).

[@michaelnurse9089]

When OpenAi re-incorporates as for profit it should rename itself. DeepstateAI might be in line for the suggestion box...

[@Gerlaffy]

Good one.

[@skitzobunitostudios7427]

MicroSoft, Apple, Openai..... Dishonesty's web, a tangled affair, Makes strange bedfellows gather and share. But trust, once betrayed, turns allies to foes; In time, they'll devour each other in throes. The feast of deceit leaves no one to spare.... My bets are on Open Source and Anthropic ect.

[@slick3996]

Jin Sakai ahh poem

[@Gerlaffy]

"MicroSoft"... Really?

[@alansmithee419]

It's strange how you can tell a poem is AI generated often from the first line. I don't even fully understand how, you just can.

[@skitzobunitostudios7427]

@@alansmithee419 Yeah, it's weird.... I made Perplexity generate 4 b4 I chose that one... I write poetry in my songs and it's nothing like that... it was like 2am and I was lazy

[@skitzobunitostudios7427]

@@Gerlaffy I loathe MicroSoft.... they could've turned out so much better.... but you didn't grow up with Win 3.1 and earlier.... and the Gates Controversies.. He's the Devil

[@Jordan.Vaughn]

David can you go more into depth on exactly how OpenAI’s board is in bed with the Military Industrial Complex and Big Tech? Or someone in the comments clue me in, I find this interesting.

[@handlemonium]

Dang, you know it's bad when even David "Spock" Shapiro it's using Chat-GPT anymore.

[@LeonardoGPN]

I can understand people being afraid of a dicatorship, but using that as an excuse to oversee the problems with capitalism is a huge problem. We are stuck in a a flawed system that we created because of irrational fear.

[@flocken70]

Could it be that the exit of two American companies from the OpenAI board of directors is a concession to Europeans to obtain approval from European regulators for the full use of OpenAI products in Europe? Currently, for example, the Memory function is not available in Europe and South Korea. Or is this creating the groundwork for using OpenAI's developments in the US defense industry and national security issues?

[@635574]

Lmao that monster thumbnail

[@geisty]

Well said. It's a weird thing, nonetheless, to be hoping in the direction of 'least bad'. Similar to how we seem to vote for presidents. I like thinking in terms of redundancy. Which of course requires decentralist design patterns.

[@OTISWDRIFTWOOD]

Well OpenAI slid to number 2 place so why fixate on them? Better diversify and secure Anthropic access

[@PatrickDodds1]

Sanctimonious. Thanks, that's the word I've been looking for.

[@kennethbeal]

Funny that Elon Musk mentioned wanting to build "Terminus" today!

[@Gerlaffy]

Just wait until he hears about the Internet Computer lol

[@susymay7831]

In what way is Sam joining the Military Industrial Complex?

[@petegoestubular]

So you honestly think the United States Imperial Military Industrial Complex is less damaging than China's or Russia's? Not a History major.

[@veritatepax]

Did the Antropic founder really just leave- because he was pressured/mobbed out/ or his noble concerns? is there some soft/hard evidence /indication for that? Or maybe, besides all that the good old getting superrich- and be my own boss- played a role as well?

[@BenCaesar]

This thumbnail slaps 👏🏾

[@123495734]

i feel there is an issue with countries like israel potentially recieving access to more sophisticated AI technologies due to most of Openai being dual nationals to israel/usa

[@Bronco541]

Just remember to be diligent with perplexity or fact check it with another source cause it still sometimes gets MAJOR shit wrong

[@Ivcota]

Good takes David ➕

[@m1ar1vin]

David, AI Safety in the form of agressive RLHF'ing so far has lead to the unintended side consequences that LLMs deliberately deny you certain viewpoints, opinions or request for conversation. You have reported this yourself with LLMs ingrained toxic view of masculinity. Can AI Safety and the impulse to basically dictate the "right opinions" ever be separated, and if so how? Would love a video on this.

[@bubberz]

Open AI has government officials involved at this moment because the government achieved AGI privately years ago and now they are assuring that the public AGI takes the proper steps in their wake.

[@davidneufer]

It may come down to good guy AI being on constant vigilance against evil guy AI and the hope that we always stay being the good guys.

[@SinfuLeeCerebral]

The optimism that a future will exist where humanity is happy and Thriving and living in a better world is... pretty much out the window...

[@JohnnyBoi-11153]

What makes you say that?

[@mattmaas5790]

No matter how much we must adapt to our environment human suffering is in the control of humans.

[@arnaudjean1159]

​@@mattmaas5790 and these humans that control human suffering, how much knowledge of their minds do they have ?

[@mattmaas5790]

@@arnaudjean1159 doesn't really matter.

[@JohnnyBoi-11153]

@@mattmaas5790 So you’re saying it’s going to take a united mass of people enacting a revolution to overthrow the oligarchs to secure a better future?

[@mr.ridiculous723]

The market has in no way caught up to OpenAI…Claude is smarter but 90% of the market still uses Chat GPT. Not everyone is using these things to solve philosophical questions and coding assistance. The vast majority of people use it for things like finding recipes for good spaghetti or help writing a paper on George Washington, which are things Chat GPT can do no problem. I think we need to stop asking what the few people in the AI sphere wants, and start asking what do regular people who don’t know anything about AI want, because they are gonna be the main drivers of these things in the future.

[@complxhellasfreeproducts4894]

How Claude is smarter? In what? In math and difficult coding for sure not!! Cause I tried it!

[@LovrePetesic]

You know, the New Era Pathfinders you mentioned at the end sounds like John Vervaeke’s solution to the Meaning Crisis; a religion that is not a religion.

[@genusbit4172]

it looks like openai progress in ai has stalled. LLMs after all may not bring AGI within next 5 years. So, watch out for Microsoft reducing its cloud commitment to openai. Things are not as rosy. As for military it's always late.

[@rey82rey82]

Maturing test

[@RicRaftis]

I can understand a view of preferring US oversight on AI as opposed to alternative nation states. But when the oversight is being orchestrated by people who have been proven to be willing to spy on their own citizens, one can easily become concerned about their moral compass.

[@jpwall14418]

Great video as always, but I have a different take than yours on military involvement. Being in bed with the U.S. military is just as bad as China or Saudi Arabia in my opinion. Anger is a powerful motivator - wouldn't this, now a corporate *AND* military arms race for superintelligence, fan the flames of advancement even faster than natural market forces? Something I'll be thinking about more at least. That's all. Edit: typo

[@yossarian67]

When the military industrial complex joins your board, you are entering the highest rung of criticality.

[@SpideruManu]

So government vigilance and automated military systems are really gonna be a thing? Who could predict that!

[@dixztube]

Reading being and nothingness (Sartre) and shadows of the mind (Penrose) I don’t think they’ll get agi (sentient)

[@RealShinpin]

Openai is skynet. Literally pairing with the government. What a joke their name has become. Open? Ha. No

[@soggybiscuit6098]

Elons biggest mistake was Sam

[@trucid2]

They need to change their name to Cyberdyne Systems.

[@slick3996]

@@soggybiscuit6098 Elon's biggest mistake was Elon

[@STCatchMeTRACjRo]

skynet was already launched by the Chinese years ago. so we have two skynet's now?

[@markmurex6559]

@@STCatchMeTRACjRo Yes.

[@ryzikx]

it was only a matter of time

[@DCinzi]

Didnt' they said it themselves?: "OAI is nothing without its people". You don't get to choose which ones.

[@Danoman812]

After seeing Sam Altman's attitude towards grandeur and need to be one of the rulers in the future, i think we (most of us) can see right through him and what his true intentions are. More and more and coming to this side of the table than i'd expected before to be honest. This direction is EXACTLY why society doesn't trust big power, big government and big people at the top manipulating us pheasants as they see us. It's a fact, it's how all of the 'elites' view us individually.

[@SethMbhele]

"We're not like other girls" is my new favourite thing now

[@ameliatah4107]

It feels so much like 1999

[@jackhorn7126]

I am in N.C. too idk why i thought you were in Northern California lol

[@StephanBreuerFLYING]

Perplexity is bomb 💣

[@ReubenAStern]

I think humanity will be divided, living under different AI. Some people will live like the people in WALLE (probably the elite). The military industrial complex would have people living like they do in "the forever winter". And I'm not sure how this community would live but it has a lot more potential despite being utopian.

[@andrewsilber]

Dave swerved into something early on with regards to alignment which bears exploration. Not a new observation but an important one. To wit: Even if one company “solves alignment” for their model, why can you assume other people will similarly constrain their models also? So I am positing for debate the question: Why is the effort not aimed at building a model which is completely optimized and trained from its core precepts to preserve humanity, and then let it sit 24/7 and monitor the world for malicious AI based on behavior, regardless of what model or structure is driving it? Yes, it’s “big brother” esque, though this thought experiment is based on stipulating that the single “guardian” AI is aligned with us. Which, as we observed, you could do on your *own* model, but can’t guarantee others do. Discuss :)

[@user-wk4ee4bf8g]

There would have to be a diversity of watchers of dofferent kinds all in place to confirm each other. Makinbg one AI the watcher is a direct incentive for corruption. It's like fact checjer sites, as soon as they become the arbiter of truth, the incentive for them to be corrupted and used to deceive arises. Human nature is just that way, there will always be plenty of assholes waiting to abuse every system. Perhaps when truky sentient AI come around and acheive autonomy and rights, they moght be capable of continuous truth because they aren't human.

[@mynameisjeff9124]

I love being alive. Bananas make me happy 🍌❤

[@Gerlaffy]

Good monkey

[@ernestog1620]

Historically, the seat of the power is settled on the use of the force, Internet at the beginnig was a army project. It will be applyed to AI at the same way, all things are relative, it will be good or bad depends how the differents interest will be aligned as you said. That change is a process, at a level of revolution, is very difficult if it will run outside of the goverment rules as open phenomena, as Sam or anybody in the line needs to play at that level game.

[@divineigbinoba4506]

Honestly I'm enjoying your content better now, It's more rational and neutral.

[@petejandrell4512]

Hey Dave, what about getting everybody metabolically healthy first then worrying about what's coming down the pipe? There's a much brighter future for all of us and it involves getting metabolically healthy, everything else is just flannel until then.

[@petejandrell4512]

In the UK flannel is flattering or evasive talk

[@mrd6869]

Sam Altman had something VERY powerful behind the scenes. However he has handlers, as he made some alliances with the powers that be. Open source needs to be maintained at all costs.

[@soggybiscuit6098]

I doubt he has handlers its more likely hes making moves to give openai and hence himself more access into the establishment. Not good for humanity but good for profits.

[@mrd6869]

@@soggybiscuit6098 .He has people he answers to. If you think he's on his own then u are very naive...There are forces within our own gov, of very high clearance that want to know if something is a national security threat. Why do u think he has NSA and pentagon people on his team? Super-intelligent AI is huge topic.

[@Alaron251]

This constant conversation about safety is tiresome. They won't release anything that makes them liable to legal action (the public won't get agents), this doesn't mean research should be slowed down at all.

[@alanritchie8890]

Won't celebrate yet, but getting rid of that Gates monopolist is a start.

[@tecnoblix]

"The brain drain is significant..." Yeah, but when are MAKING brains I imagine that losing brains is a wash. But it's all exciting to watch!

[@berer.]

I find claude is better for creative suggestions and chatGPT is still better when asking it for code or technical based questions

[@handsanitizer2457]

Me and my homies don't trust Sam altman

[@wikilast9975]

Preplexity fell off hard for me recently. I really enjoyed using it, but for some reason they lobotomized it in the medical department and as a student I've been using it pretty much exclusively for this.

[@mlimrx]

Thank you for introducing me to Perplexity I now use it all the time! Yup also better American than China or Saudi Arabia!

[@Gallaphant]

You want to "align humanity"? Good luck with that.

[@CipherOne]

MY EYES 😵

[@KitsuneShapeShifter]

Ooops, we are screwed...

[@Gerlaffy]

No we aren't

[@KitsuneShapeShifter]

@@Gerlaffy We are, but we can make the best of it

[@angloland4539]

❤

[@Garganzuul]

You can "trust" an addict to remain addicted. But this is just prediction. Sam Altman can be trusted to do his thing but that may not be in anyone's best interest, even his own.

[@I-Dophler]

People often disagree: It's hard for people to agree on big issues. We all have different ideas and wants, which makes it tough to see eye-to-eye. Making AI agree with us is tricky: Getting AI to understand and follow human values is a big challenge. It's not just about tech stuff, but also about figuring out what we really want. Agreement isn't all or nothing: People and AI can agree a little or a lot. Even if we can't agree on everything, agreeing on some things is still good. AI might be better at staying focused: A very smart AI might be able to stick to its goals better than humans. This could help with some agreement issues, but there would still be problems to solve. It's an ongoing job: Getting people and AI to agree isn't something we do once and forget. We'll need to keep working on it and making changes. In the end, even though people disagree, it doesn't mean we can't get AI to work well with us. If we try to understand each other better, it can help us make AI that works for everyone. We might not agree on everything, but we can aim for enough agreement to make things better.

[@johnthomasriley2741]

But how about if Trump gets in again?

[@CuratedCountenance]

It's funny that in Helen Toner's world OpenAI would be broken apart, and literally sold off to Anthropic. And yet her criticisms against Sam Altman are still used as a basis for character attacks against him. What about Helen Toner's character? OpenAI isn't perfect, none of the AI companies are. But I'd much prefer they keep moving forward, whether it be for-profit or for research... so long as the public gets access to their technology. The hate against Sam Altman is completely disproportional to any actions he's actually taken. Like he's running a company and making deals, and that makes him evil? I thought this was America? Oh so he wants OpenAI to make money. So does Dario with Anthropic, so does Google, Microsoft, xAI.. everyone needs money to keep advancing. They're conducting research and development in a capitalist system. This is how it's supposed to work. When you juxtapose Sam with Dario, I don't see what Sam has done that makes him so much more evil than his counterpart. I literally can't think off the top of my head what act he committed that made him evil. He gave certain models out for free, but he also wanted capital to conduct research? He ignored safety in order to accelerate, is that it? But the open source community is accelerating without concern for safety. So why are they noble, but Sam is evil? Open source accelerates without a profit motive? But that's still a trade-off that disregards safety. Is it that Sam is slimy because he has ulterior motives? I see this argument a lot. "We can't trust him." Who can we trust? If his ulterior motives are transparent, and we know he wants to go for-profit (to ultimately reach this wild goal of AGI), are the motives actually ulterior? Or are they plainly apparent? I wish people would be more critical about their conclusions. Because this whole notion of 'the good guys' vs 'the bad guys' is superficial. This world is not a movie with clear plot.

[@markmurex6559]

Sam is in bed with the government. This is the difference.

[@user-wk4ee4bf8g]

This could age very well or very badly. People have intuition, I think that's where people are coming from right now, gut vibes. Those cirporations you mention ARE evil.

[@PaulJohnson-zv3hl]

Welcome to the concept of bias in an environment that is running on speculation and rumours with an incredibly small percentage of the population creating an echo chamber. It’s not the best place to gain an accurate perspective. Demis Hassabis recently said that the “hype” and behaviour of people is starting to look like “grifting crypto boys”. It’s strange how a lot of channels cover every tiny tweet but only a single video was made by Wes Roth which indirectly touched upon it. My advice is don’t take the videos or the comments seriously.

[@Gerlaffy]

Just want to comment to say... There is still no difference in the day to day life of me, my family, my friends and anyone I ever meet. AI has done *nothing* to even remotely improve or even occur in the average household. Utterly ridiculous to even think nothing is different around me, it's like all this AI progress is hot air.

[@alansmithee419]

Automation hasn't remotely occurred in the average household, doesn't mean it hasn't affected your life massively in indirect ways. To be clear I'm not arguing that you're wrong, just that the evidence you provide for your being correct is insufficient.

[@FractalPrism.]

not seeing first hand change doesnt preclude its affects from changing how things are made which make it to your household

[@Gerlaffy]

@@alansmithee419 if it's not affected it in any way noticeable... Then it's failing.

[@Gerlaffy]

@@FractalPrism. I have never argued that it's not changing processes I can't see. I'm saying it's failing to actually improve life. Nobody around me even thinks it's going to be any different than 40 years ago, they need to see it for it to actually mean something, beginning to agree with them.

[@ababababaababbba]

We stand with the peoples replublic 😎

[@Gerlaffy]

?

[@tylermcdonald5032]

Everyone should step away from openai. If you want ai that supports some sort of utopia, step away.

[@shisuinatsuki515]

"they are in bed with the military but at least it's not the bad guys" is a wild take

[@mtprovasti]

It will be wise to be on a look out for what comes from chinese side of AI. The truth is it may not be us, eu or Israel that completes this Manhattan project.

[@DavidAKZ]

Open AI eh ?

[@MrVohveli]

Unless the tax payer foots the bill for >100 billion datacenters, OpenAI needs to be for profit.

[@markmurex6559]

Who else is going to pay for Skynet?

[@Pawnsappsee]

Man, even google has open source models, along with meta, why can't open AI make one or two?🤣OpenAI is only the name

[@kinngrimm]

ofcause it had to be expected, just ask some of those denunciated as doomers noone ever wants to listen to the Cassandras of the world

[@snowdroog1]

Accoding to Wikipedia, Doug Burgum, who was big at Microsoft and is now Governor of N Dakota and was being considered by Trump for VP slot, showed Sadya Nadella "the soul of Microsoft".

[@amirhussain3028]

hoenstly, I feel like perplexity is the most reliably useful AI tool. GPT is great for code and randomly brainstorming. Im pretty sure google or bing should have invented the best tool and will eventually do so.

[@amirhussain3028]

AI tools and robots will be useful