Having the feature on a menu somewhere would be fine. The problem is the confluence of new features now becoming possible, and companies no longer building software for their users but as vehicles to push some agenda. Now we’re seeing this in action.
Part of the model of products like Adobe's Creative Suite [1] is that they are always adding new features -- and if you want people to keep renewing their subscription you want them to know about new features so they feel like they are getting more out of their product.
Trouble is using a product like that is like walking out of the Moscone Center and getting harassed by mentally ill people and addicts or like creating an account in Tumblr and getting five solicitations for pig butchering and NFT scams in DM in the first week -- you boot up the product, spend 20 seconds looking at the splash screen, then you have to clear five dialog boxes that you might not have time to deal with right now. Sometimes I open up a product because I have to do a task I have to do but don't really want to do and feeling a lot of stress and I just don't need to deal with any bullshit when I am under the gun.
I've seen Adobe trying gentler methods to point out new features in Lightroom, such as a filter that can automatically weed out photos where people have their eyes closed. It takes a lot of UX work to do that though.
Personally I'd like it a lot better if the nagging started after I finished a task, if I was feeling satisfied with the product and now relieved that the task is over that's a moment when I'd be receptive to learning more about the product.
[1] And also a lot of "free" software, it's not just money-grubbing, but the model of always rolling updates.
This is the fundamental problem and it has nothing to do with AI. Just look at the recent iOs 26 release. I am not convinced that any of the actual functional changes warranted a release or that they needed to be released at that point if a new release was needed. New software to justify new phones.
You get lots of features but performance takes a second seat. And sadly, I feel it works. I feel most would balk at paying a monthly subscription if only performance related improvements were made
Why do we need OS changes though? Well practically we don’t. But the platform owners all want to move new hardware so they need to shovel features in, which we could just completely ignore, except that they’ll abandon you to the wolves for security patches, which is about the only “new” thing we do need, if you’re not on the latest couple releases. And as for hardware, eventually you need new hardware and drivers only get created for current and future OS releases.
So the end result is we’re being led on a wild goose chase of trend-chasing shitty UI changes, adware, and performance-killing crap we don’t need, purely because we can’t run the old hardware forever, and even when we can keep the old hardware going, we can’t safely run old software for lack of patches.
But now I’ve got several bugs (and I’m on last years flagship), liquid glass is ugly until you change a guy of settings, and I find myself accidentally triggering something (usually Siri) and being annoyed more.
And this is why the subscription model just doesn’t make sense for most businesses. I pay for a newspaper subscription because there is literally a brand new newspaper each day. A magazine subscription yields an entirely new set of articles every month. I pay for subscription access to data that is continuously updated. The subscription model makes sense for a product that is created anew on a regular basis. It doesn’t make sense for most software companies that are producing static software. What they are calling ‘subscriptions’ are really just rentals for their static products that get minimal surface changes to justify the ongoing rent charge. I’d much rather just pay a flat fee for the static software and upgrade it when I’m ready for the new features.
This is especially irritating when, say, you set up a new phone and the app treats you as if you've never used it before.
Really my complaint is anything that covers up content; if instead of popping up a popover Firefox just took 75px above or below the page to show me something I’d complain about lot less — but if I had my way anything unwanted that covers unwanted content should bust down the whole c-suite to working in an Amazon warehouse. (I could trust those folks to deliver stuff with an e-bike but don’t want anybody with bad judgement like that driving a car or truck!)
All companies push an agenda all the time, and their agenda always is: market dominance, profitability, monopoly and rent extraction, rinse and repeat into other markets, power maximization for their owners and executives.
The freak stampede of all these tech giants to shove AI down everybody's throat just shows that they perceive the technology as having huge potential to advance the above agenda, for themselves, or for their competitors at their detriment.
I'll bear that in mind the next time I'm getting a haircut. How do you think Bob's Barbers is going to achieve all of that?
https://www.thebignewsletter.com/about
> The Problem: America is in a monopoly crisis. A monopoly is, at its core, a private government that sets the terms, services, and wages in a market, like how Mark Zuckerberg structures discourse in social networking. Every monopoly is a mini-dictatorship over a market. And today, there are monopolies everywhere. They are in big markets, like search engines, medicine, cable, and shipping. They are also in small ones, like mail sorting software and cheerleading. Over 75% of American industries are more consolidated today than they were decades ago.
> Unregulated monopolies cause a lot of problems. They raise prices, lower wages, and move money from rural areas to a few gilded cities. Dominant firms don’t focus on competing, they focus on corrupting our politics to protect their market power. Monopolies are also brittle, and tend to put all their eggs in one basket, which results in shortages. There is a reason everyone hates monopolies, and why we’ve hated them for hundreds of years.
https://blogs.cornell.edu/info2040/2021/09/17/graph-theory-o... (Food consolidation)
https://followthemoney.com/infographic-the-u-s-media-is-cont... (Media consolidation)
https://www.kearney.com/industry/energy/article/how-utilitie... (US electric utilities)
https://aglawjournal.wp.drake.edu/wp-content/uploads/sites/6... [pdf] (Agriculture consolidation)
https://www.visualcapitalist.com/interactive-major-tech-acqu... (Big Tech consolidation)
I think part of the Mozilla problem is that they are based in San Francisco which puts them in touch with people from Facebook and Google and OpenAI every frickin' day and they are just so seeped in the FOMO Dilemma [1] that they can't hear the objection to NFT and AI features that users, particularly Firefox users, hate. [2]
I'd really like to see Mozilla move anywhere but the bay area, whether that is Dublin or Denver. When you aren't hanging out with "big tech" people at lunch and after work and when you have to get in a frickin' airplane to meet with those people you might start to "think different" and get some empathy for users and produce a better product and be a viable business as opposed to another out-of-touch and unaccountable NGO.
[1] Clayton Christensen pointed out in The Innovator's Dilemma that companies like Kodak and Xerox die because they are focused on the needs of their current customers who could care less about the new shiny that can't satisfy their needs now but will be superior in say 15 years. Now we have The FOMO Dilemma which is best illustrated by Windows 8 which went in a bold direction (tabletization) that users were completely indifferent to: firms now introduce things that their existing customers hate because they read The Innovator's Dilemma and don't want to wind up like Xerox.
[2] we use Firefox because we hate that corporate garbage.
(1) Fully fund Firefox or an alternative browser (with a 100% open source commitment and verifiable builds so we know the people who get ideas like chatcontrol can't slip something bad in)
(2) Pass a law to the effect: "Violate DNT and the c-suite goes to jail and the company pays 200% of yearly revenue"
(3) same for having a cookie banner
Seems like maybe forking it in an agreeable way, and funding an EU crew to do the needful with the goal of upstreaming as much as possible.
I don't have insight into EU investments but that would provide a lot of bang for their euros.
(Also, point of order: Opera was always based in Norway, which is not a member of the European Union.)
Regulation.
some weeks if its slow he may struggle to make his rent for his apartment; he doesn't have time or capacity to engage in serious rent-seeking behavior.
but hair cut chains like Supercuts are absolutely engaging in shady behavior all the time, like games with how solons rent chairs or employing questionably legal trafficked workers.
and FYI turns out that Supercuts a wholly owned subsidiary of the Regis Corporation, who absolutely acquires other companies and plays all sorts of shady corporate games, including branching into other markets and monopoly efforts.
But if users really wanted agenda-free products and services, then those would win right? At least according to free market theory.
Not once in the history of tech “the free market” has succeeded in preventing big corps or investors with lots of money from doing something they want.
2. AI could be the next technology revolution
3. If we get on the AI bandwagon now we're getting in on the ground floor
4. If we don't get on the AI bandwagon now we risk being left behind
5. Now that we've invested into AI we need to make sure we're seeing return on our investment
6. Our users don't seem to understand what AI could possibly do so we should remind them so that they use the feature
7. Our users aren't opting in to the features we're offering so we should opt them in automatically
Like any other 'big, unproven bet' everyone is rushing in. See also: 'stories' making their way into everything (Instagram, Facebook, Telegram, etc.), vertical short-form videos (TikTok, Reels, Shorts, etc). The difference here is that the companies put literally tens or hundreds of billions of dollars into it so, for many, if AI fails and the money is wasted it could be an existential threat for entire departments or companies. nvidia is such a huge percentage of the entire US economy that if the AI accelerator market collapses it's going to wipe out something like ten percent of GDP.
So yeah, I get why companies are doing this; it's an actual 'slippery slope' that they fell into where they don't see any way out but to keep going and hope that it works out for them somehow, for some reason.
Similar to how I read about a bar in the UK that has an intentional Faraday cage to encourage people to interact with people in the real world.
That's the core issue. No one wants to fail early or fail fast anymore. It's "lets stick to our guns and push this thing hard and far until it actually starts working for us."
Sometimes the time just isn't right for a particular technology. You put it out there, try for a little bit, and if it fails, it fails. Move on.
You don't keep investing in your failure while telling your users "You think you don't want this, but trust us, you actually do."
I think there are more mundane (and IMO realistic) explanations than assuming that this is some kind of weird power move by all of software. I have a hard time believing that Salesforce and Adobe want to advance an agenda other than selling product and giving their C-suite nice bonuses.
I think you can explain a lot of this as:
1. Executives (CEOs, CTOs, VPs, whatever) got convinced that AI is the new growth thing
2. AI costs a _lot_ of money relative to most product enhancements, so there's an inherent need to justify that expense.
3. All of the unwanted and pushy features are a way of creating metrics that justify the expense of AI for the C-suite.
4. It takes time for users to effectively say "We didn't want this," and in the meantime a whole host of engineers, engineering managers, and product managers have gotten promoted and/or better gigs because they could say "we added AI" to their product.
There's also a herd effect among competing products that tends to make these things go in waves.
If we didn't have pervasive telemetry, we also wouldn't have these obnoxious nudges; UX teams would get their feedback from QA testing and focus groups, and leave the end users in peace.
https://www.acquired.fm Acquired podcast does long (2 4-hour episodes on Google) episodes on various companies, mostly tech but recently Trader Joe's
Google Plus was 100% hubris. “If we build our version of Facebook, it course everyone will flock to it.”
Doing nothing while a competitor gains steam would've been hubris.
Maybe I’m wrong and internally they knew they had a major uphill battle, but I don’t think so. So many of the choices they made were needlessly user hostile (e.g. real name requirements) that it seems like they assumed it would be a given that people would want to use it. When they later realized their error they tried to cram it down everyone’s throats with stuff like YouTube comments only working from Google Plus accounts.
I think you're wrong with probably the same confidence you think you're not wrong. :)
At most, I'd say they didn't expect it to be as hard as it proved to be.
I totally agree that Google just didn't get it right, but all the things you describe, to me, fall under a mix of "they had to try", and "it was working for Facebook" (but also having to differentiate from Facebook at the same time, eg with circles).
(Disclaimer, I guess) I was working for Facebook when the whole Google Plus thing happened, and Facebook definitely saw it as a serious threat. I don't at all recall Facebook folks laughing it off as Google hubris, more like it was a long shot, but Google wasn't to be ignored.
Upvote for you regardless, because I think it's a solid take and an engaging comment.
I'm trying to remember all of the crap integrations with the likes of Youtube that were pushed. Just, screw that stuff. And quit trying to make yet another new messenger app!
Firstly, that whole account-unification thing where YouTube accounts were getting merged with Google[+] logins. That rubbed me the wrong way.
Then the Google+ promotional stuff all talked about how you could use "Circles" to silo posts to different "circles" of friends. It sounded very complicated and I was worried that I'd publish something snarky to the wrong group of friends :)
I wonder how many others had the same concern? Given that Steve Yegge accidentally published one of his rants to the public that was meant purely for internal Google consumption (I think that was on G+ ...?) that might have been a legit thing to be wary of.
There was also the very minor annoyance of G+ taking over the + operator in Google search (previously you could say +keyword instead of "keyword" to force literal search), but I don't think that would have swayed me against joining.
If you can’t solve the chicken and egg problem of engagement then nothing else really matters.
I never used any of its collaboration features, just looked at them. I did use it as a friendly-for-non-geeks version of IRC for a group of people that lived in three separate cities as a virtual watch party for LOST. And for that, it was spectacular even if it was painfully slow on a netbook (so was everything else, but it was cheap and light and worked).
The thing is that a Google Mail early invitee could collaborate with everybody else via the pre-existing standard of SMTP email. They felt special because they got a new web UI, told their friends about it, generated hype, which then made the invites feel even more special, etc...
Google Wave had no existing standard to leverage, making it 100.00% useless if you couldn't invite EVERYBODY you needed to collaborate with. But you couldn't! You weren't allowed! They had to wait for an invite. Days? Weeks? Months? Years!? Who knows!
There was a snowball's chance in hell that this marketing approach could possibly work for a collaboration tool like Google Wave, but Google knew better. They knew better than every journalist that pointed this obvious flaw out. They knew better than every blog post, Slashdot commenter, etc...
It was one of the most spectacular failures caused by self-important hubris that I've ever seen in any industry.
https://www.newschoolers.com/news/read/Got-Google-Wave-Invit...
https://bolesblogs.com/2009/10/13/get-your-google-wave-invit...
https://googlewaveinvites.weebly.com/
And so on: https://www.google.com/search?q=%22google+wave%22+%22invites...
I don't personally care if a product includes AI, it's the pushiness of it that's annoying.
That, and the inordinate amount of effort being devoted to it. It's just hilarious at this point that Microsoft, for example, is moving heaven and earth to put AI into everything office, and yet Excel still automatically converts random things into dates (the "ability" to turn it off they added a few years ago only works half the time, and only affects csv imports) with no ability to disable it.
I mean, c'mon, its literally called the fucking windows key and it doesn't work. As per standard Microsoft it's a feature that worked perfectly on all versions before cortana (their last "ai assistant" type push), i wonder what new core functionalities of their product they're going to fuck up and never fix.
Windows as an OS really kind of peaked around Windows 7 IMO... though I do like the previews on the taskbar, that's about the only advancement since that I appreciate at all... besides WSL2(g) that is. I used to joke that Windows was my favorite Linux distro, now I just don't want it near me. Even my SO would rather be off of it.
Microsoft could have made Windows privacy respecting, continued investing in WSL, baked PowerToys into the OS, etc. and actually made one hell of a workhorse operating system that could rival the mac for developer mindshare. They could partner with Google and/or Samsung and make some deep Android integration to rival Apple's ecosystem of products. Make Windows+Android just as seamless and convenient as mac + iOS.
Instead they opted for forced online accounts, invasive telemetry, and ads in the OS instead of actually trying to keep and win over the very enthusiasts that help ensure their product gets chosen in the enterprise world where they make their cash.
Now they're going to scrap the concept of Windows as something you interact with directly all together and make it "Agentic" whatever the hell that means.
I don't think their bet is going to pay off, especially if the bubble crashes. I think it will be one of the biggest blunders and mistakes that Microsoft will have made.
I agree with you in principle, but in practice these two are currently inextricable; if there's AI in the product, then it will be pushed / impossible to turn off / take resources away from actual product improvement.
Hopefully after the pop rather than shoving it in our face they can return to advertising at us to use the things, and the things needing to prove themselves to get to real sales, rather than corporations getting 10% stock pumps in a day based on statistics about how "used" their AI stuff is while they don't tell the market how few people actually chose to use their AI stuff rather than just becoming a metric when it was pushed on them.
I can only hope they won't change it back at the next update (already happened once).
Just to push their annoying google assistant
Although I never saw anybody reporting it was actually useful, it's tasteful, accessible, and completely out of your way until you need it.
You can disable AI in Google products.
E.g. in Gmail: go to Settings (the gear icon), click See all settings, navigate to the General tab, scroll down to find Smart features and personalization and uncheck the checkbox.
> Important: By default, smart feature settings are off if you live in: The European Economic Area, Japan, Switzerland, United Kingdom
(same source as in grandparent comment).
(I desperately want to disable the AI summaries of email threads, but I don't want to give up the extra spam filtering benefit of having the smart features enabled)
Google now "helpfully" decides that you must want a summary of literally every file you open in Drive, which is extra annoying because the summary box causes the UI to move around after the document is opened. The other day I was looking at my company's next year's benefits PDFs and Gemini decided that when I opened the medical benefits paperwork that the thing I would care about is that I can get an ID card with an online account... not the various plan deductibles or anything useful like that.
I turned off the "smart" features and the only thing that changed is that the nag box still pops up and shifts the UI around, but now there's a button that asks if you want a summary instead of generating it automatically.
And the worst thing is not only is it being pushed, it is being pushed at the expense of UI/UX. No, Google, I don't need 'help to write' or 'to summarize this document'. I can read and write just fine. And the worst thing of all is that you can't turn it off because they'll just move it around every other week.
I want to choose the extensions that go into my browser. I don't even use the browser's credential manager, and I've gotten to a point where I'm just not sure anything is actually getting better.
I will say that the Gemini answers at the top of Google searches are hit or miss, and I do appreciate that they're there. That said, I'm a bit mixed as the actual search results beyond that seem to be getting worse overall. I don't know if it's my own bias, but when the Gemini answer is insufficient, it feels like the search results are just plain off from what I'm looking for.
ai features in the right context are truly awesome, but the engagement hacking is getting old.
Maybe I'll ask Gemini to write one...
LLM's are a product that want to data collect and get trained by a huge amount of inputs, with upvotes and downvotes to calibrate their quality of output, with the hope that they will eventually become good enough to replace the very people they trained them.
The best part is, we're conditioned to treat those products as if they are forces of nature. An inevitability that, like a tornado, is approaching us. As if they're not the byproduct of humans.
If we consider that, then we the users get the shorter end of the stick, and we only keep moving forward with it because we've been sold to the idea that whatever lies at the peak is a net positive for everyone.
That, or we just don't care about the end result. Both are bad in their own way.
You're completely correct, that's fair criticism. The excitement made me skip the basics. Here's a quick breakdown:
What it does: It's a new optimization algorithm that finds exceptionally good solutions to the MAX-CUT problem (and others) very quickly.
What is MAX-CUT: It's a classic NP-hard problem where you split a graph's nodes into two groups to maximize the number of edges between the groups. It's fundamental in computer science and has applications in circuit design, statistical physics, and machine learning.
How it works (The "Grav" part): It treats parameters like particles in a gravitational field. The "loss" creates an attractive force, but I've added a quantum potential that creates a repulsive force, preventing collapse into local minima. The adaptive engine balances these forces dynamically.
Comparison: The script in the post beats the 0.878... approximation guarantee of the famous Goemans-Williamson algorithm on small, dense graphs. It's not just another gradient optimizer; it's designed for complex, noisy landscapes where Adam and others plateau.
I've updated the README with a "Technical Background" section. Thanks for the push—it's much better now.
Sounds like a return to "Clippy the paperclip" or the dog from the ill fated Microsoft Bob [1] that insisted on always popping up every five to ten minutes with something like: "I see you may be entering a ????, would you like to make it a ??? ???".
And I 'member that you could program it from VBA somehow. Think via OLE, but I was a kid back in the Clippy era.
Which meant you could use it in Internet Explorer but not anywhere else. But it did make for some interesting web pages. I built a custom one with the mascot of the university I was attending at the time. It was, let's say, some peak 1990s internet. (Never shipped it to anyone, just had it internally.)
That took some non-trivial web searching. "Microsoft" "Agent" and most of the other keywords are pretty well covered by a few million other web pages by now.
ActiveX and OLE... technologies ahead of their time, eh. VB, VBA, Internet Explorer, standalone VBScript, C/C++ - didn't matter, it all was (trivially) interoperable.
I don’t know what version of Gemini they’re stuffing into Google products, but sheets, docs, and colab/data science agent are all bad experiences.
If you aren’t putting something comparable to good paid models into your product then don’t bother putting that feature out.
Once you train your users that your ai is half baked junk they’re not coming back to waste their time with it. It’s 10x as frustrating than regular product failures.
As far as I can tell Gemini in gsuite can do nothing other than summarise text and regular LLM q&a (but with Gemini’s perennially sad, apologetic persona)
If the nagging didn't work would companies keep doing it? Someone's KPIs must be increasing for them to keep doing it.
It's fucking Clippy all over again