Rebuff5007

I believe in silicon valley terms, this is called "moving fast and breaking things"

Its a bit sad to me that the tech community here on HN doesn't seem to take any responsibility for all this.

Surely some non-trivial percentage of the commenters / lurkers that are proud to talk about their mono-repo or their favorite react library had some part in the fact that millions think the covid vaccine has 5G.

My negativity for Thinking Machines is precisely because others (such as yourself) already consider it a success.

I know this is the leading narrative, but I actually disagree.

Apple has a wonderful set of products without any form of generative AI, and those products continues to exist. Yes there is opportunity to add some fancy natural-language based search / control, but that seems like relatively low hanging fruit compared to the moat they have and defend.

Will adding gen ai natively to apple products have people non-trivially change the way they use iphones or macs? Probably not. So there is literally no rush here.

I vote for a campaign to make sure every single entrant has this meme on their phone.

I dont understand, why would they make a super-intelligence group instead of something more ambitious like a super-super intelligence group or supreme-mega-intelligence group? Zuck is clearly not thinking this one through.

There is so much that is sad and infuriating about this.

How can a stranger have so much hate towards someone who's content they want to consume? Why are meta engineers suggesting suing their own company instead of pointing to someone to walk through a basic appeals process? The obvious, why do so many things in the world not work without a smartphone and a google/facebook account with internet access? Why isn't anyone doing anything about any of this? PMs at facebook? Congressmen/women?

The system is not working.

Im not sure this is a great example... yes the infrastructure of posting and applying to jobs has to go, but the cost of recruitment in this world would actually be much higher... you likely need more people and more resources to recruit a single employee.

In other words, there is a lot more spam in the world. Efficiencies in hiring that implicitly existed until today may no longer exist because anyone and their mother can generate a professional-looking cover letter or personal web page or w/e.

> clearly in many cases superhuman

In what cases is it superhuman exactly? And what humans are you comparing against?

I'd bet that for any discipline you chose, one could find an expert in that field that can trick any of today's post-gpt3 ais.

I'd argue its a bad concept in any spaces that involve teams of people working together and deliverables that enter the real world.

Really? Open-AI says PhD intelligence is just around the corner!

You (a SWE) using google maps is a bit different than a cab driving relying on google maps. The former is fine, the latter could be concerning.

I'd argue that coding assistants for a SWE hinders their ability to build subject matter expertise.

I find it assuming that the might of the American government -- in trying to take a bunch of data offline -- is being resisted by a digital "militia" of hobbyist archivers and non profits.

Theres something that about this that just rings second amendment. Personally I think the concept of civilians having weapons to be a check on a nation state is absurd, but in this case it feels pretty empowering.

In RL literature this is generally called "curriculum learning".

The curriculum is usually modeled as some form of reward function to steer learning, or sometimes by environment configuration (e.g. learn to walk on a normal surface before a slippery surface).

Tangential, but this has gotten me thinking...

I used to wonder how the hundreds of thousands of employees that work in Big Oil or Big Pharma could tolerate all the terrible things their company does... e.g. the opioid epidemic. The naive optimist in me never thought that the tech industry would ever be that bad.

Now, as someone thats been in the industry for 10+ years and working adjacent to LLMs, this is all so depressing. The hype has gotten out of control. We are spending hundreds of billions of dollars on things that simply are not making life better for the majority of people.

Can somehow explain to me how they can simply not respect copyright and get away with it? Also is this a uniquely open-ai problem, or also true of the other llm makers?

I'm tearing my hair out... how is the solution here not just better data privacy laws? Doesn't that solve all the issues, both domestic and international?

I'm so curious about how the internal dev teams feeling about all this scaling. Four cities across 3 states -- surely there are differences in road signs, lane markings, emergency procedures, etc. Let alone the sheer volume of data of doing hundreds of thousands of miles ever week!

Massive kudos to them if they are able to do all this without things being aflame on the inside...

My definition of "works" includes the fact that a self-driving car will never drive into a parked fire truck, or many other things i've seen tesla FSD do.

Tesla testing and developing FSD with normal consumer drivers frankly seems criminal. Test drivers for AV companies get advanced driver training, need to filed detailed reports about the cars response to various driving scenarios, and generally are paid to be as attentive as possible. The fact that any old tech-bro or un-assuming old lady can buy this thing and be on their phone when the car could potentially turn into oncoming traffic is mind boggling.

"You either die a hero or you live long enough to become the villain"

The whole industry is full of snake oil salesmen. Do you really think this [2] is not a hype bubble waiting to be turned into a tell-all docuseries?

They have produced something impactful sure, but I don't think their product is a tenth as valuable or "intelligent" as they are claiming it to be. There are too many domain-specific tools that are far superior to whatever pseudo-intelligence you get from chatgpt... what proof do we have that any industry has found material value from this chatbot?

[2] https://x.com/deedydas/status/1841670760705949746

I for one can never get over the fact that Mira Murati was not laughed out of the room when she said -- with a straight face -- that GPT4 had high school level intelligence and the non-existent GPT5 will have PHD-level intelligence [1].

IMO -- this is not a serious company with serious people building an important long-lived product. This is a group of snake oil salesmen that are in the middle of the greatest grift of their careers. That, and some AI researchers that are probably enjoying limitless gpus.

[1] https://www.timesnownews.com/technology-science/next-gen-cha...

> - I do not know if there is a clean, ethical way to "run the social media"

My hand-wavy proposal:

1. there needs to be something akin to a constitution where all players involved (users of social media, social media companies) can express some shared set of values. For example kids shouldn't get depressed, data should be private, widely spread information should be reasonably accurate.

2. There needs to be a few institutions with enough power and checks and balances to be able to steer the system towards these values.

Excuse me, what?! How is this cargo cult science? I literally just described the scientific method. Here it is described on Wikipedia:

> The process in the scientific method involves making conjectures (hypothetical explanations), deriving predictions from the hypotheses as logical consequences, and then carrying out experiments or empirical observations based on those predictions.

The other questions you added are not bad questions to ask if you are practical and resource-constrained, but this is besides the point.

https://en.wikipedia.org/wiki/Scientific_method

To produce scientific work, one needs to follow the scientific method. This involves stating a hypothesis, designing an experiment that would test this hypothesis, conducting the experiment with controls, and analyzing the data w.r.t to the hypothesis being tested.

Access to past works is only useful in informing what is a good hypothesis worth testing. And data is only useful when generated by an experiment that is testing for causality (see [1]).

[1] https://pyimagesearch.com/2023/11/27/a-brief-introduction-to...

If apple was a young startup, struggling to get market share, or struggling to turn a profit, I can understand why they might want a cut of profits here.

But they literally have more cash than anyone in the world... why bother with this?!

There's a group in UT Austin that has spend lots of time trying to model and analyze such dynamics: https://www.cs.utexas.edu/~aim/

> ads online are much more energy-intensive than billboards.

Is this quote from the article remotely true or even verifiable?

In fairness, Mira Murati said GPT-4 is only high school level [1]. Maybe it takes PhD level to understand basic shapes?

[1] https://www.ccn.com/news/technology/openais-gpt-5-phd-level-...