I would also add that search has already moved elsewhere.
Less and less people are using search engines to shop, ex:Amazon makes >$57B a year from search ads, but also look at Temu and Shein which are mostly glorified product search platforms.
No one is searching for "funny videos" when you can just open Instagram and Tiktok.
The only real unique thing that search engines can do is queries that are not directly commercial (e.g. education, information seeking, etc.) and competition is insanely intense (w/ ChatGPT, Perplexity, etc) there.
Honestly, I haven't used either of ChatGPT or Perplexity seriously. They haven't performed particularly well when I tested them and dun-dun-dun in my uses Gemini has been growing on me. And another odd thing at the moment is that Google's search has somehow become better at giving me the results I'm looking for and DDG is giving a lot of annoying crap.
Basically when I search for API of a specific function or package docs on DDG I end up with page after page of people blogging about using them and the actual docs don't show up. So I add "!g" and the same crap is there, but the link to reference will be somewhere among the first page of results (although Goggle usually has a link to an old stale version of the docs).
Do you have specific examples of this behavior that I can look into? Also, curious if you've tried our Assist function (comes up automatically for some searches or click Assist under search box) or duck.ai for stuff like that?
I didn't realize that the context would require such so much memory. Is this KV caches? It would seem like a big advantage if this memory requirement could be reduced.
> I cannot be the first person to think about such possibilities
Differentiable Rendering [1] is the closest thing to what you are describing.
And yes, people have been working on this for the same reason you outline, it is more data/compute efficient and hence should generalize better.
But also:
> While cool, this also seems utterly wasteful. Video games offer known "analytical" solutions for the interactions that the model provides as a "statistical approximation", so to say.
A bit of the same debate as people calling LLMs a "blurry JPEG of the web" and hence useless.
Yes this is a statistical approximation to an analytical problem... but that's a very reductive framing to what is going on.
To find the symbolic/analytical solution here would require to constrain the problem greatly: not all things on the screen have a differentiable representation, for example complex simulations might involve some kind of custom internal loop/simulation.
You waste compute to get a solution that can just be trained on billions of unlabeled (synthetic) examples, and then generalize to previously unseen prompts/environments.
Training code is only useful to people in academia, and the closest thing to "code you can modify" are open weights.
People are framing this as if it was an open-source hierarchy, with "actual" open-source requiring all training code to be shared. This is not obvious to me, as I'm not asking people that share open-source libraries to also share the tools they used to develop them. I'm also not asking them to share all the design documents/architecture discussion behind this software. It's sufficient that I can take the end result and reshape it in any way I desire.
This is coming from an LLM practitioner that finetunes models for a living; and this constant debate about open-source vs open-weights seems like a huge distraction vs the impact open-sourcing something like Llama has... this is truly a Linux-like moment. (at a much smaller scale of course, for now at least)
I dunno — if an open source project required, say, a proprietary compiler, that would diminish its open source-ness. But I agree it's not totally comparable, since the weights are not particularly analogous to machine code. We probably need a new term. Open Weights.
"The world probably has never been as peaceful as in the last 50 years or so."
So all the defense ministers in europe for example, who urge us to get ready for real war, are paranoid then?
25 years ago it was mostly peaceful, with a steady downward trend since then.
And many people, including government officials, see now a war between china and US as inevitable. That is different from my youth, which was mainly optimistic, with the big players negotiating and demilitarizing.
In the past people who were mostly disinterested in politics and news simply didn’t watch/read the news. Now the news finds them and they don’t have enough of an understanding of civics, economics, and history or the emotional coping mechanisms to know what to do with it.
It’s not that they were fully ignorant of the news in the past, but people used to get most of their “takes” and opinions from like, a weekly editorial column or Sunday paper or something that would sort of bubble up through various cultural filtering mechanisms. By the time it got to your typical 17 year old it has already gone through a “processing mechanism” where it’s been analyzed, contextualized, better understood. They’re not getting the messy information in raw form.
Now they get the raw information as it happens, and often bubbling up from agenda driven activists instead of disinterested analysts. Consequently, everything has a heavy emotional charge and is specifically geared to short circuit the rational mechanisms for understanding things and hit you in the lizard brain. The lizard brain is good for quick thinking and rapid reactions, but it’s not very good at emotional regulation, priority setting, or any of the other stuff that’s conducive to mental health.
> The world probably has never been as peaceful as in the last 50 years or so.
True, but we are not talking about the last 50 years, we are talking about the future, for example, the next 5 years. People don't commit suicide if they think their future will be OK.
> Same goes for access to drinkable water, food, decent shelter, gender equality, freedoms, technology, etc.
- Item 2 and 3 ignore recent rises in food prices and housing prices.
- For the U.S., item 4 and 5 have been reduced in many places by various recent laws and judicial decisions (unless you are super rich), and this will probably continue for the near future.
- The last item - 6 - technology. I think for a non-tech-person standpoint, it really appears like technology primarily has ...
A) ... stagnated. Phones--the tech most people use daily--haven't done anything new and exciting for some time. The Internet hasn't given the non-tech public-at-large the next big thing after social media, either.
B) ... primarily become interested in putting monthy fees, ads and tracking on everything one does, and
That may be correct -- depending on the degree to which you agree with Pinker's (I assume) interpretation of statistics. But it probably does not reflect the reality of the people contemplating or committing suicide. They may also not care if things are _objectively_ better when things subjectively suck for them.
But, very few people people commit suicide purely in response to their despair concerning the state of the world. There are almost always deeper mental health issues involved. And as strange as it sounds, suicide goes in and out of fashion as well. In general, it's very difficult to drill down on causality of suicide rates.
It seems like you’re vaguely alluding to global averages when most commenters are lamenting the state of their own countries, communities, daily lives - a categorical error.
Scale & locality can’t be hand-waved away.
"the world" probably doesn't mean that much to a 25-30 year old American's personal morale; I don't see why it should
Like, I'm sure someone in Flint a few years ago wouldn't begrudge more people having access to drinkable water but they're obviously going to be massively unhappy about having to consider the possibility of no longer having access to it themselves.
Ironically, suicide rates are typically fairly low among people dealing with actual existential problems. Depression is not very prevalent until all basic needs are taken care of.
Imagine how all the millions of warfighters over the last few decades in the middle east feel about this hysteria. And pretty much the best 98% of Americans could do was a flag sticker and a yawn.
But their phone says this is different, so they believe it. Mission Accomplished.
Water with micro plastics and availability under threat due to climate change. Some places are already feeling it.
Food that’s hyperpalatable, full of sugar, and devoid of nutrition. Case-in-point, the obesity crisis.
Sure technology is great but the internet has fucked our brains. The tech industry repeatedly creates hype bubbles like crypto to make a buck. And don’t even get me started about AI.
Gender equality is far from resolved for both men and women. Women still have to deal with harassment and men are now expected to fulfill contradictory roles of being both a progressive feminist and a traditional masculine man.
Sure things are great when you oversimplify and ignore the bad parts.
About as many people as die in road traffic incidents every year in the USA, basically completely preventable, and nobody (media, congress, president, policial candidates) are kicking up a fuss about that. 47k estimated in 2020: https://www.statista.com/statistics/192575/road-traffic-fata...
The parent is intentionally conflating the last 2 years with the prior 5 decades. Right now there is a war going on, of a magnitude not seen since WW1 and WW2.
Yeah, things are better than ever in most metrics, but this is also the issue with metrics. Those same metrics do not cover two current wars that involve nuclear powers and the risk of these two wars becoming global conflicts. Those metrics do not cover increasing atomization of human societies and the impact of increased atomization on the human social animal. Those metrics do not cover human exposure to sunlight, clean air, clean water, and other health factors. Likewise, humans have a tendency to focus on risks and negatives rather than celebrating wins as a product of evolutionary pressure.
Yes, the person you’re replying to reminds me of those guys on Twitter who selectively quote stats to try and prove that people are doomers and things are actually good.
There are so many things that statistics don’t cover it’s unreal.
One thing that especially comes to mind is that many of the illusions of society have been shattered by the younger generations. Things like believing you should be loyal to your employer because they’ll return the favor.
> "Things like believing you should be loyal to your employer because they’ll return the favor."
Nobody's believed that since, I'd say, the 1970s-1980s. Even Japan, famous for its culture of lifetime employment, phased that out in the 1990s. It's ridiculous to suggest that that's a new phenomenon.
Surely you jest. Coasting and resting-and-vesting have been around longer than most of the people reading this site have been alive. That some influencers and hack journalists rebranded it to quiet quitting and marketed it as the trendy thing to do does not remotely make it a new phenomenon.
No i don’t jest. Coasting and rest-and-vest have a completely different vibe than quiet quitting. The former is more like “fuck it”, whereas the latter is more like “fuck you”.
Burying your head in the sand is definitely more comfortable, as is repeating stats. But if you live in the same reality as everyone else and are paying any attention whatsoever, seems hard to believe that you wouldn't take notice of the Russian war in Ukraine, Israel's attack on Gaza, tensions heating up between Israel & its neighbors, Switzerland holding a vote on whether or not to remain a neutral company in the event of a war, Russia & China constantly attacking countries' computer & physical infrastructure, etc.
Your platitudes do nothing to make up for the fact that wealth inequality is worse than it has ever been many times over. The cruel and unusual experience of such exploitation and social contortions faced by those at the bottom will always undermine your pathetic attempts to legitimate such ignorance.
The quality of life of the poorest members of society is also much higher than it has ever been, indeed, much higher than historical kings and lords.
Unfortunately, people do tend to be happier if they're miserable among other miserable people, than if they were living comfortably among people with cooler things than them.
Your first point conflicts with your second. Maybe work on that.
It is very well understood across all relevant metrics that younger people are less economically secure than their parents. Coupled with extreme and growing wealth inequality, this is bound to be a political problem whether you like it or not. And, these suicide numbers suggest it may be a serious psychological problem as well. Your platitudes don't change that.
Feudalist societies were not organized by wealth but rather by roles that people were essentially born into. These roles were not achieved but rather understood as basically granted by god. Just as kings had divine rights, people had god-given qualities and understood themselves as having inherent and inherited positions in society, which effectively eliminated all of the social pressure and unusual contortions experienced by subjects under capitalism, even at the level of peasants. So-called fairness as we understand it today was not even a coherent concept at the societal level. Peasants owned nothing but also had far more freedoms and access to resources, which were eventually deemed property as humans entered into capitalism, which forced peasants into capitalist production as workers.
To understand the differences between feudalism and capitalism, you must also understand the conflict between the feudalist aristocracy (god-given) and the capitalist bourgeoise (industrious accumulators of capital). This conict was really the source of the shift to capitalism. The aristocracy could not compete with the capitalists as the capitalists promoted a story of mobility by accumulating capital. Of course, however, only a few can achieve that, and they basically still argue that their abilities are god-given. But the claim about abilities and hard work paying off vs. simple i herited permanent roles results in a much more damaging psychology.
And telling young people that they're crazy for noticing these things isn't going to help either. I'm not old, but I can't call myself young anymore, and I understand now that if I live to a reasonable life expectancy, I'm probably going to see some shit. I can't imagine tacking on a decade or three to that; I used to envy children born after me for the advances they'd live to see, but now I feel nothing but sorrow for them, and I'm so relieved that I never had any.
Never mind those looming clouds or that biting breeze that just picked up, it's been sunny for a whole week -- it's sure to last forever this time! /s
Those last 50 years were a side effect from one of two global powers catastrophically collapsing. The tensions that drive modern history didn't dissolve when that happened, they were just tabled as numerous contending powers with their own visions of the world reorganized and regrouped. Maybe we can all diplomatically navigate our way past the worst outcomes along the way, but the next 50 years simply aren't going to look like the last 50.
> The world probably has never been as peaceful as in the last 50 years or so.
oh jeez, say that to the people of Palestine or the Congo. people who make these blanket generalizations are either tone deaf or come from a place of privilege.
Anyone who says that isn't aware of how many wars there used to be. See[1] and look at the countries and regions involved, and how in this century it's down to a very very few regions. Since WWII, Europe has been incredibly peaceful after centuries of invasions, border conflicts, revolutions, empires rising and falling.
Pointing to one thing going wrong so you can call people 'privileged' doesn't change it. Nuclear Mutually Assured Destruction (MAD), NATO, the EU, the USA, the generally calm dissolution of the British/French/Spanish empires, the end of new lands to find, has stabilised a lot.
cost effective in what sense? groq doesn't achieve high efficiency, only low latency. but that's not done in a cost-effective way. compare sambanova achieving the same performance with 8 chips instead of 568, and with higher precision.
Most important, even ignoring latency, is throughput (tokens) per $$$. And according to their own benchmark [1] (famous last words :)), they're quite cost efficient.
This (innovator's dilemma / too afraid of disrupting your own ads business model) is the most common explanation folks are giving for this, but seems to be some sort of post-rationalization of why such a large company full of competent researchers/engineers would drop the ball this hard.
My read (having seen some of this on the inside), is that it was a mix of being too worried about safety issues (OMG, the chatbot occasionally says something offensive!) and being too complacent (too comfortable with incremental changes in Search, no appetite for launching an entirely new type of product / doing something really out there). There are many ways to monetize a chatbot, OpenAI for example is raking billions in subscription fees.
Google gets much more scrutiny then smaller companies so it's understandable to be worried. Pretty much any small mistake of theirs turns into clickbait on here and the other tech news sites and you get hundreds of comments about how evil Big Tech is. Of course it's their own fault that their PR hews negative so frequently but still it's understandable why they were so shy.
Sydney when initially released was much less censored and the vast majority of responses online were positive, "this is hilarious/cool", not "OMG Sydney should be banned!".
It's understandable that people at Google are worried because it's likely very unpleasant to see critical articles and tweets about something you did. But that isn't really bad for Google's business in any of the ways that losing to someone on AI would be.
Google is constantly being sued for nearly everything they do. They create a Chrome Incognito mode like Firefox's private browsing mode and they get sued. They start restricting App permissions on Android, sued. Adding a feature where Google maps lets you select the location of your next appointment as a destination in a single click, sued (that's leveraging your calendar monopoly to improve your map app).
Google has it's hands in so many fields that any change they make that disrupts the status-quo brings down antitrust investigations and lawsuits.
That's the reason why Firefox and Safari dropping support for 3rd party cookies gets a yawn from regulators while Google gets pinned between the CMA wanting to slow down or stop 3rd party cookies deprecation to prevent disrupting the ads market and the ICO wanting Google to drop support yesterday.
This is not about bad press or people feeling bad about news articles. Google has been hit by billion dollar fines in the past and has become hesitant to do anything.
Where smaller companies can take the "Elon Musk" route and just pay fines and settle lawsuits as just the cost of doing business, Google has become an unwieldy juggernaut unable to move out of fear of people complaining and taking another pound of flesh. To be clear, I don't agree with a strategy of ignoring inconvenient regulations, but Google's excess of caution has severely limited their ability to innovate. But given previous judgements against Google, I can't exactly say that they're wrong to do so. Even Google can only pay so many multi-billion dollar fines before they have to close shop, and I can't exactly say the world would be better off if that happened.
That's true for google, sure. But what about individual workers and managers at google?
You can push things forward hard, battle the many stakeholders all of whom want their thing at the top of the search results page, get a load of extra headcount to make a robust and scalable user-facing system, join an on-call rota and get called at 2am, engage in a bunch of ethically questionable behaviour skirting the border between fair use and copyright infringement, hire and manage loads of data labellers in low-income countries who get paid a pittance, battle the internal doubters who think Google Assistant shows chatbots are a joke and users don't want it, and battle the internal fearmongers who think your ML system is going to call black people monkeys, and at the end of it maybe it's great or maybe it ends up an embarrassment that gets withdrawn, like Tay.
Or you can publish some academic papers. Maybe do some work improving the automatic transcription for youtube, or translation for google translate. Finish work at 3pm on a Friday, and have plenty of time to enjoy your $400k salary.
>There are many ways to monetize a chatbot, OpenAI for example is raking billions in subscription fees.
Compared to Google, OpenAI's billions is peanuts, while costing a fortune to generate. GPT-4 doesn't seem profitable (if it was, would they need to throttle it?)
There could be an opposite avenue: ad-free Google Premium subscription with AI chat as a crown jewel. An ultimate opportunity to diversify from ad revenue.
The low operating margin of serving a GPT-4 scale model sounds like a compelling explanation for why Google stayed out of it.
But then why did Microsoft put its money behind it? Alphabet's revenue is around $300bn, and Microsoft's is around $210bn which is lower but it is the same order of magnitude.
Monetizing a chatbot is one thing. Beating revenues every year when you are already making 300b a year is a whole different ball game
There must be tens of execs who understand this but their payout depends on keeping status quo
So fast it finally made virtual environments usable for me.
But it's not (yet) a full replacement for conda, e.g. it won't install things outside of Python packages
This shouldn't be surprised, e.g. the model != the product. The same way GPT4o behaves differently than the ChatGPT product when using GPT4o.
reply