Minori's recent activity
-
Comment on US Centers for Disease Control and Prevention to end all monkey research in ~science
-
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentSatisfiers vs optimizers as someone on Tildes said! I have this compulsion to find "the best" thing too, and I've successfully fought it by deciding to buy a cheap or mediocre version of something...Satisfiers vs optimizers as someone on Tildes said!
I have this compulsion to find "the best" thing too, and I've successfully fought it by deciding to buy a cheap or mediocre version of something then replace it if that's insufficient.
If I'm shopping for expensive or long-term things, I'll do more research, but my stress really dropped when I realized the cost of switching is usually very low!
As long as the cheap bulb doesn't set my home on fire, I can always buy a better one later. Or if it turns out I don't need rain pants as often as I expected, I've saved myself a lot of time and money!
-
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentActually, copyright came out of a British system for government censorship of printing press works: https://en.wikipedia.org/wiki/Copyright#Conception Before copyright, intellectual and artistic...Copyright was made precisely because large companies would take some new invention and outscale the inventor with no compensation. No copyright would benefit big companies more.
Actually, copyright came out of a British system for government censorship of printing press works: https://en.wikipedia.org/wiki/Copyright#Conception
Before copyright, intellectual and artistic works were considered truly free and open. Copyright exists exclusively to create "intellectual property" and protect profit.
I agree it'd be a massive improvement to shorten copyright terms.
-
Comment on GPT-5 has come a long way in mathematics in ~tech
Minori Link ParentHere's a a blog post about why they tend to get 3.10 - 3.9 wrong, especially in English. I personally wouldn't trust them, especially when calculators are so ubiquitous.Here's a a blog post about why they tend to get
3.10 - 3.9wrong, especially in English. I personally wouldn't trust them, especially when calculators are so ubiquitous. -
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentI made a post here which incidentally discusses the emotional tone that companies are giving LLMs. You're of course correct that there's no way to generate text that is completely devoid of...I made a post here which incidentally discusses the emotional tone that companies are giving LLMs. You're of course correct that there's no way to generate text that is completely devoid of emotion (since neutrality itself is an emotional state).
However, companies have intentionally increased sycophancy because it drives engagement, and we all know how that works out in the attention economy...
Companies can't make their models completely dull, but they could and should limit the sycophancy and glazing that encourages pscyhosis in users.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentThey could do math. Everyone knew the issue was coming, but nobody wants to draw the short straw. Even unions have continued to put their pension debts on the credit card and beg someone else to...They could do math. Everyone knew the issue was coming, but nobody wants to draw the short straw.
Even unions have continued to put their pension debts on the credit card and beg someone else to pay the tab.
-
Comment on Boeing's Starliner to return to ISS in cargo-only capacity for now in ~space
Minori Link ParentThey're doing a lot of hiring because of new funding from the US "Department of War"! Their workforce is also heavily unionized. It's generally a good place to work; better than its non-unionized...They're doing a lot of hiring because of new funding from the US "Department of War"!
Their workforce is also heavily unionized. It's generally a good place to work; better than its non-unionized competitors if you care about work-life balance.
-
Comment on Against 'Metroidbrania': a landscape of knowledge games in ~games
Minori Link ParentI use roguelite when I mean roguelite because the average reader still sees roguelike. Even in speech, it's similar enough that most people won't notice. I think the difference between the two...I use roguelite when I mean roguelite because the average reader still sees roguelike. Even in speech, it's similar enough that most people won't notice. I think the difference between the two words is useful to know, but it's not worth being prescriptive over.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentOr, CEOs like Bezos could have their compensation taxed more heavily. It's actually really really easy to tax stock grants or discourage things like stock buybacks. Simply raising the top end of...but someone like Bezos should have at some point in his life have to sell (or borrow against) some of the value he has access to by owning a big part of Amazon, otherwise wealth equality and redistribution will have slim odds of happening.
Or, CEOs like Bezos could have their compensation taxed more heavily. It's actually really really easy to tax stock grants or discourage things like stock buybacks. Simply raising the top end of taxes and instituting an inheritance tax would cover the vast majority of wealth.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentThis is certainly a take, considering the state of many unions. Syndicalism largely died as an ideology precisely because it turns out that labor can rent-seek and don't necessarily benefit...Labor, almost by definition, can't rent-seek. If you remove a landlord from his rentals, they still function quite well. If you remove labor from a company, it ceases to exist.
This is certainly a take, considering the state of many unions. Syndicalism largely died as an ideology precisely because it turns out that labor can rent-seek and don't necessarily benefit society without corporations balancing them.
As a great example, see police unions in the US. Blue states with higher rates of police union membership pay cops more than teachers while red states with lower union membership pay teachers much more. The awful policing in many cities is a direct function of police union rent seeking. I don't often cite Jacobin, but I largely agree with them here. The main thing they miss is that any union, especially in the public sector, can become a rent seeker that worsens society. Absent a corporation or competition, unions are just as prone to bias and cronyism.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentCitation needed. Investments have almost always been taxed separately because the economics are totally different. In most cases, long-term capital gains are subject to progressive brackets....Citation needed. Investments have almost always been taxed separately because the economics are totally different. In most cases, long-term capital gains are subject to progressive brackets. Short-term gains are still taxed as income in the US.
You're correct that capital gains were originally taxed as income in the US, but other countries like the UK didn't tax investments till after WW2 as more people became wealthy enough to invest.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentInheritance or estate taxes are separate from recurring wealth taxes. In fact, they solve most of the issues inherent to wealth taxes! They provide an easy reference point for valuations; it's...Inheritance or estate taxes are separate from recurring wealth taxes. In fact, they solve most of the issues inherent to wealth taxes! They provide an easy reference point for valuations; it's pretty hard to manipulate time of death.
-
Comment on A new era of intelligence with Gemini 3 in ~tech
Minori Link ParentThis isn't true. Most pensions are failing because they were setup as pyramid schemes which assumed there would always be more young workers paying in than old pensioners making withdrawals. It's...I'll point out that state (well all) pensions are primarily doomed because of neoliberals gutting them, rather than any intrinsic failure of the pension system itself.
This isn't true. Most pensions are failing because they were setup as pyramid schemes which assumed there would always be more young workers paying in than old pensioners making withdrawals. It's no coincidence that the failure of many pension schemes neatly aligns with population pyramids inverting.
They are also underfunded. Neither workers nor corporations have contributed enough funds for them to pay out.
-
Comment on LLMs are bullshitters. But that doesn't mean they're not useful. in ~tech
Minori (edited )LinkNote: This is a personal essay by Matt Ranger, Kagi’s head of ML
In 1986, Harry Frankfurt wrote On Bullshit. He differentiates a lying from bullshitting:
-
Lying means you have a concept of what is true, and you’re choosing to misrepresent it.
-
Bullshitting means you’re attempting to persuade without caring for what the truth is.
Fine tuning makes some kind of text more statistically likely and other kinds of text less so.
Changing the probabilities also means that Improving probability of a behavior is likely to change the probability of another, different behavior.
For example, the fully finetuned Gemini 2.5 will correct user inputs that are wrong.
But correcting the user also means the model is now more likely to gaslight the user when the model is confidently wrong:
LLMs are Sophists
Historically, bullshitting had another name: sophistry. The sophists were highly educated people who helped others attain their goals by working their rhetoric, in exchange for money.
In that historical conception, you would go to a philosopher for life advice. Questions like “How can I know if I’m living my life well?” you would want to pose to a philosopher.
On the other hand, you go to a sophist to solve problems. Questions like “How can I convince my boss to promote me?” would go to a Sophist.
If I use a LLM to help me find a certain page in a document, or sanity check this post while writing it, I don’t care “why” the LLM did it. I just care that it found that page or caught obvious mistakes in my writing faster than I could have.
I don’t think I need to list the large number of tasks where LLMs can save humans time, if used well.
But remember that LLMs are bullshitters: you can use LLMs to get incredible gains in how fast you can do tasks like research, writing code, etc. assuming that you are doing it mindfully with the pitfalls in mind
By all means, use LLMs where they are useful tools: tasks where you can verify the output, where speed matters more than perfection, where the stakes of being wrong are low.
But don’t naively trust a system that freaks out at the inexistence of the seahorse emoji to complete critical tasks without your supervision.
Your therapist or partner should not be a bullshitter
You should not go to an LLM for emotional conversations. An LLM is capable of emitting text that is a facsimile of what an emotional conversation sounds like. An LLM is not capable of emotions. Models outputting statistically probable text cannot and should not be a replacement for human connection.
. . .
Interacting with sycophantic models like this reduces willingness to repair interpersonal conflict and increases users’ conviction of being in the right.#Sycophancy is good for the spreadsheet
On a similar note, we know that sycophantic model behavior worsens users’ mental health. But sycophancy also tends to be rated more favorably by users overall.
So companies that optimize user retention are likely to end up encouraging sycophancy. A user with mental health issues is better than no users in a financial spreadsheet.
Conclusion
LLMs are useful, if used well. Don’t to trust them too much, and don’t use products that assume hands-off trust of an LLM as a core part of the user experience.
And finally: make sure that technology you use is actually working in your best interest. If it’s working in someone else’s interest, make sure you understand where your interests are at odds with that.
-
-
LLMs are bullshitters. But that doesn't mean they're not useful.
19 votes -
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentYour examples are funny to me because I actually view them mostly opposite. LED bulbs can vary massively; do they flash, have the right color temp, last a long time, etc. RAM is pretty easy to...Your examples are funny to me because I actually view them mostly opposite. LED bulbs can vary massively; do they flash, have the right color temp, last a long time, etc. RAM is pretty easy to shop for as long as you pick the right criteria around timings and clock speed. USB-C on the other hand is super easy if you know what protocol you want and whether you want a USB-IF certified cable.
I agree with you on aesthetics being particularly difficult to describe in advance. It's hard for me to trust an LLM will be able to select the same filters I value and identify the ideal product. Although, "fine" is good enough for plenty of products.
-
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentThis is very similar to what Amazon has tried with Alexa voice shopping. The difference is they own Alexa and have an opportunity to shoehorn ads and their own products.This is very similar to what Amazon has tried with Alexa voice shopping. The difference is they own Alexa and have an opportunity to shoehorn ads and their own products.
-
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentThe Amazon phone app is notoriously bad. There's a reason I only use the website. They make money off the search and UI getting worse. More searches means more opportunities to show ads!The Amazon phone app is notoriously bad. There's a reason I only use the website. They make money off the search and UI getting worse. More searches means more opportunities to show ads!
-
Comment on The DoorDash problem: How AI browsers are a huge threat to Amazon in ~tech
Minori Link ParentPatents are particularly unfair. They're abused to hell and back. I honestly believe we'd be better off without copyright and patents. Ideally, they'd be restricted to 5-10 years to encourage...Patents are particularly unfair. They're abused to hell and back. I honestly believe we'd be better off without copyright and patents. Ideally, they'd be restricted to 5-10 years to encourage creativity.
-
Comment on GPT-5 has come a long way in mathematics in ~tech
Minori Link ParentThat's more true for things like proofs, but LLMs are not calculators. They're trained on plenty of math, yet they frequently get basic arithmetic wrong. They don't have any deeper "understanding"...That's more true for things like proofs, but LLMs are not calculators. They're trained on plenty of math, yet they frequently get basic arithmetic wrong. They don't have any deeper "understanding" of math. If they did, they'd be able to consistently and correctly apply basic mathematical operations.
I'm quite confident that far more than 3-5 monkeys get used in important trials and experiments.