A few hours ago I wanted to know the difference between a typhoon and a hurricane. Here is the answer: they're the same thing, just in a different ocean. I clicked three or four articles and had to read for a few minutes to get to that answer.
It's even worse with videos. It takes 10 minutes to answer the simplest questions, because that's the ideal length for monetisation.
I just googled “what is the difference between typhoon and hurricane”
Top result is from the Red Cross, with this blurb conveniently extracted so I don’t even need to click into the article unless I’m curious more deeply:
If it's above the North Atlantic, central North Pacific or eastern North Pacific oceans (Florida, Caribbean Islands, Texas, Hawaii, etc.), we call it a hurricane. If it hovers over the Northwest Pacific Ocean (usually East Asia), we call it a typhoon.
""" The only difference between a hurricane and a typhoon is the location where the storm occurs. If a storm is above the North Atlantic, central North Pacific, or eastern North Pacific oceans, it's called a hurricane. If it hovers over the Northwest Pacific Ocean (usually East Asia), it's called a typhoon. """
A Hurricane has a ~1000hp early Rolls Royce Merlin engine and a top speed of about 350mph; a Typhoon has a 2000hp+ Napier Sabre engine and a top speed of over 400mph.
The Typhoon was actually intended to be the replacement for the Hurricane, but challenges in the high altitude interceptor role led to the Typhoon taking on a more fighter-bomber role as the war went on.
Edit: at full afterburner, it burns fuel at a heat output of 235 MW (expressed as horsepower: 315140)
> "Let’s start with some research from Hook Agency, which claims the best content length for SEO in 2023 is between 1,760 and 2,400 words"
The source, however loathsome, is probably correct. The front pages of the internet have become saturated with garbage ultra-long-form articles... not even "articles," really, more like ramblings. And it's Google's fault.
<wall of text useless content>
<start of bulleted ingredients list>
* Rosemary - I remember the first time smelling fresh rosemary, it was the spring of 1989 and I was in my dear aunties kitchen...
* 1 boneless skinless chicken brest - Chicken breasts are a common ingredient in many dishes throughout the world. This hearty chunk of meat contains a high amount of protein and is quite versatile...
Google has been enshitifiying the web for a good while now, and I used to work there! I worked hard to make the core product more efficient but ultimately your internal goals just override the external ecosystem. And with so much power google was able to exert a shifting force on the web.
The classic example is mega-long intros for recipes. But if you look into it, the primary reason is to be able to include 10 ads on a page instead of 2.
That was the difference you meant to highlight, right?
Answer: apple sauce and gelatin
I use FF mainly on desktop and mobile, and there's probably a great opportunity for Mozilla to build an offline, privacy-first summarization model.
[0] https://help.kagi.com/kagi/getting-started/faqs.html#i-have-...
For the record I am a happy user, not affiliated.
Build as much of the model as you can in the cloud, run inference locally and push results back is probably the cost optimal way to run this stuff at scale.
I assume so too, and given that, it's incredibly frustrating (but not at all surprising) that they require users to use Chrome to be able to use the feature.
I think Kagi might do this? Any other good solutions?
[0] https://kagi.com/summarizer/index.html?url=https%3A%2F%2Fwww...
Also, all this wailing and gnashing of teeth is really quite annoying. Look at you people embarrassing yourselves.
With this, Google controls how content you read is summarized, they could introduce biases, intentionally or not, and I'm not sure we should trust such an actor for this.
It's one thing to get recommendations in a place where recommendations are appropriate, and the dangers of the feedback loop you are talking about. It's another thing entirely to actively push it in other contexts.
"I know you are searching for programming topics, but might you be interested in this political outrage instead?"
And if I wanted to find more in depth material to support my firmly held beliefs I just had to go to the right bookstore.
How do you know it works well? How do you know it misses nothing? How do you know it's not subtly biased (intentionally or not) in a way you would not notice, or you would think it's good enough?
> I can read articles against my view and still get only what I want. My world view will become increasingly polarized, and it's already quite biased.
Yes, of course we all do, but it's another thing to involve a third party in this process.
I wish I had your experience. My YouTube recommendations can definitely be said to be “my fault” in some sense of the word, but wherever the blame lies they’re still mostly terrible. YouTube will watch me skip a video for a whole week, but keep showing it to me “just in case.” I’d rather it were just a dumb search at this point.
I might feel more kindly toward YouTube recommendations if they did this for me.
This was always the case.
That's an individual solution. I'm interested in a world that's as good as possible not only for me, but for the other people too. Most people around still use Google. They might find this new feature useful and start using it, and possibly suffer this bias. Not only they may get manipulated individually, but it can have broader consequences, even to those avoiding the feature. If Facebook can influence which president get elected in a country, it will also affect people avoiding facebook (sometime for this very reason).
Avoiding stuff for oneself is a first step but it's hardly a solution, these big corporations have a non negligible effect on the world, still with you as a member.
We'll have to start to raise awareness around these things, in addition to all the subjects on which we already need awareness.
But same old same old I guess, Google already puts you in a bubble, just even more concerning I guess. It was manipulating us, now it's another, new, potential manipulation angle.
You won’t find info about any merely “controversial” idea on Google, it’s all censored, it’s all disinformation already. Try even finding memes, you generally won’t find them. Try finding info that doesn’t paint some politicians in a good light, you won’t find it, even though it exists with the exact keywords you type in and the crawler returns a success for that page.
So what do you mean, while you already trust Google to give its bias to the current world?
I will Google it and report back…
People were Hunters before. Were searching for food. We do not need now to hunt. We had to hunt for information, we do not need now.
We rely more and more on services and corporations. We choose the jest one, which becomes monopoly after some time. After some longer period of time monopolystic corporation becomes monster.
Circle of life.
We may not want to use chatgpt or other bots, but eventually we will. Common people use the easiest route. They will decide, what is popular, and what will be used.
We've got enough problems with people reading a headline and running with it. I hate to think of google summarizing some scientific study - particularly some "pop" scientific study - and have everyone confidently reciting it as though it's been conclusively proven without making any effort to look at the source.
It's like when people talk about Google search results without realizing their search/browsing history has put them in a particular bubble. There's no guarantee any two people's result ranking will be the same which can affect their understanding or perspective on a topic.
They're literally featuring a wrong ChatGPT answer as an answer to this question: https://www.google.com/search?q=country+in+africa+that+start...
Maybe de-index receipes talking about how this soup was my grandma’s preferred soup. It’s not so hard, you still have the full power on search engines, Google! For at least 2 more full months! Do it!
You are the problem, Google.
Looks like my inbox is at 250,000 emails. Wonder if search degrades with inbox size. (Same with drive?)
It's not actually a direct correlation, but enough people take that as gospel that they will pad out an article to make it super long just for the SEO value.
Similar to recipe websites giving their entire life story before a recipe.
Ironically Google is the main cause of articles getting really long and full of fluff.
The Dead Internet theory will soon be indisputable.
And if you opt out google will eventually just push you down.
Google is ahead of the curve in how ready they are for LLMs!
Results from these models are orders of magnitude more expensive than traditional search. Maybe they have some benefits of scale where they can cache and serve up the same summaries to many people?
But this doesn't change that this is a useful feature.
PS: That being said, I have started developing a distaste for google search. I have been served so much SPAM lately, that I am now starting to associate that search box with low-quality content.
This is a failure mode for the information economy.
Google's behavior in regard to informations reminds me a lot of the "torrent leeches" of the past: siphon everything you can, give nothing back, ..., profit. But at a huge scale with an impact on society far more than "a few more bytes on a personal hard drive".
Most people do not produce content to be left unread.
I'm not really worried about bias, since from my experience, summarization software will inherit the bias of the creator. Sure, it's behind generative AI, which can introduce more bias but it's not like the original article is hidden. No one is forced to use it.
For example, start by skimming and scanning an article and you may be able to pull out all the salient points and satisfy your reading goal or choose to go deeper with your analysis depending on time and interest and what you are trying to accomplish:
https://advice.writing.utoronto.ca/researching/skim-and-scan...
For those who don't like when Google summarizes content for you but like the idea of getting a link preview with a summary provided by the website owner via meta tags before navigating to a third-party website - I've built a product called Linkz.ai [0] - that allows website owners to install link preview popups with 1 line of code.
[0] https://linkz.ai
There are studies saying overall human attention span is shrinking. Anecdotally, I know people in their 20s who use TikTok daily, and they're no longer able to watch a movie without losing focus.
Squeezing all aspects of our digital life, including learning new skills or reading the news, into a bite-sized "summary," is not going to lead to a better-informed and more productive society.
I'm almost 40 and haven't been able to do this since the smartphone arrived; it's not Tiktok, it's just that the phone is right there, and movies are slow-paced (arty thrillers) or monotonous (lengthy CGI fight scenes).
Doesn't this generation have an absurdly long attention span compared to the generation of the 30 minute TV show with three 2-3 minute commercial breaks?
And news orgs are going to keep coming under assault cause they perform a similar function to search engine robots by surfacing new info. As soon as some one copies that info, the value of the info is 0.
There is already data saying most content created is never consumed by anyone just cause of how much copying and duplication happens.
- clickbait
- hyperbolic
- gossip
- propaganda
- unoriginal
- unimportant
- CBC: Propagandic, unoriginal, unimportant
- Global News: Clickbait, gossip, unoriginal, unimportant
- Toronto Star: Hyperbolic, propagandic
- The Globe and Mail: Unoriginal, unimportant
- The National Post / Postmedia network: Hyperbolic, propagandic
- CTV News: Clickbait, gossip, unoriginal, unimportant
Anything I missed?
I have found the best strategy is to take the sum of multiple sources to get a slightly less awful whole.
Luckily there are AI tools out there do this.
[1] It's off by default these days, you'll have to turn on the Summary service under service settings.
That's all I need for such a query. In comparison any text written by a "generative AI" is mostly noise.
We're all unique and unpredictable. In a given article, a random sentence or point considered throwaway by most people and algorithms could turn out to be meaningful to me.
I've been trained not to even trust the basic facts they pull of out content like movie showtimes or whatever. Once you see it wrong a few times you realize it's folly not to keep the responsibility of finding information yourself.
I find these tools useful for articles these days because everything is clickbait and full of bloat and nonsense. Even if it's just to do a first past to figure out if I want to read the full article.
But I do understand your point, I feel this is very important ' a random sentence or point considered throwaway by most people and algorithms could turn out to be meaningful to me.', i've had this issue with book summary services, i've read a book and took completely different meanings or found insight in certain paragraphs that were completely glossed over in book summarys that just changed it into generic sounding nonsense. We need to connect the dots ourselfs and to do that, I believe is not to have someone elses summary of the situation.
However, for everything else a tool like this is useful, there is too much noise in the world, we need tools to filter it and help us understand what is relevant and what is not.
Having Google summarize a too-long recipe (a problem they created, by the way) into just the simple ingredients and instructions is NOT "deciding your view of the world".
The beginning could just be "see this related product or service." Which won't mess with the content. Perhaps that's fine.
I'd argue though that any summarizing tool inherently has a bias. It must choose to ignore certain details and make decisions about what to highlight.
As we understand LLM's more and the stuff that summarizes folks controlling them will be able to make those decisions and the money and power behind that will absolutely abuse it.
That's not even considering the effects this would have on journalism and writing in general if most of it gets summarized.
On a mass scale, maybe 0.2% of people would double check the source page. The rest would be unknowingly influenced to eat "healthier" by Google.
Yes, this seems unlikely. But so has much in the last 5 years.
Why are we deliberately charging head-first into social man-in-the-middle attacks? We already don't trust each other enough. LLMs lie, and lie often. Why should we trust them for anything?
Would you be OK asking a LLM what food is safe for an infant, or a pet dog? Without checking the source?
I have no interest in AI summarizing anything at all for me. Not because of AI, but because I have no interest in anyone or anything summarizing for me. Too much is lost.
This is what I crave for in good articles and books. There's some big insight just thrown in in the middle of the sentence because that pattern is already obvious to the author. It's also what AI-generated content is currently lacking because it sticks to the task at hand (but I believe it will change).
Also, your critique applies just as much to human-generated summaries.
My issue with it is that I don't want my browser to send home what I read and view although I guess that ship has sailed some time ago.
To me it’s simply a noisy discussion with little value. Don’t use it but let’s not waste time debating it. No one’s mind is going to change.
I’d love nothing more than for an algorithm to filter out the noise and extract any information and facts. Learning whose opinion I trust on specific topics would also be useful.
Of course, others will tell me why this doesn’t work for them and they want the added noise. Seriously, the entire point is that some people will want the product while others don’t.
I’m not trying to change your mind, and I’m not trying to stop anyone else from the discussion. It’s simply not for me, in general.
It sure is.
If you had told people in 2019 that a computer program could summarize large topics and write cogent output you would have been placed in the asylum.
It can be. Summarizing process can be tweaked very slightly to give more space to certain POVs, make them sound more reliable than others via word choice etc. The subtlety can make it undetectable, yet it can have strong effects when applied on mass scale.
I don't know how this will look like, but it is a very powerful technology to sway the public opinion one way or another.
