schwarzrules

The only advantage I could anticipate is I still hit session limits with Opus 5.5. My usage shows I'm on-track reach my weekly reset with room to spare, but yesterday I ran into a session limit. I switched down to Sonnet 5 for the next session, but performance benefit of Sonnet 5.5 is a compelling alternative for managing session limits.

Just going to use this thread to also add a nit: I hate that Microsoft made all the OEMs add a Copilot key to all their laptop keyboards. It's been Ctrl forever, and suddently they throw this net-negative key onto the keyboard. I have decades of habit of using the right side Ctrl key for varying shortcut strokes, and then when I use it now it loads up an unproductive Copilot experience that is awful. <end rant>

New Yorker had a good write up on this many moons ago: https://www.newyorker.com/magazine/2009/08/10/ticketmaster-l... (non-paywalled: https://archive.ph/5ILdG

I understand all the chatter about LLMs hallucinating, or making assumptions, or not being able to understand or provide the more human/emotional element of health care.

But the question I ask myself is: is this better than the alternative? if I wasn't asking ChatGPT, where would I go to get help?

The answers I can anticipate are: questionably trustworthy web content; an overconfident friend who may have read questionably trustworthy web content; my mom who is referencing health recommendations from 1972. And as best I can imagine, LLMs are going to likely to provide health advice that's as good but likely better than any of those alternatives.

With that said, I acknowledge that people are likely more inclined to trust ChatGPT more like a licensed medical provider, at which point the comparison may become somewhat more murky, especially with higher severity health concerns.

Summary using Kagi Summarizer. Disclaimer, this summary uses LLMs, so the summary may, in fact, be bullshit.

Title: LLMs are bullshitters. But that doesn't mean they're not useful | Kagi Blog

The article "LLMs are bullshitters. But that doesn't mean they're not useful" by Matt Ranger argues that Large Language Models (LLMs) are fundamentally "bullshitters" because they prioritize generating statistically probable text over factual accuracy. Drawing a parallel to Harry Frankfurt's definition of bullshitting, Ranger explains that LLMs predict the next word without regard for truth. This characteristic is inherent in their training process, which involves predicting text sequences and then fine-tuning their behavior. While LLMs can produce impressive outputs, they are prone to errors and can even "gaslight" users when confidently wrong, as demonstrated by examples like Gemini 2.5 Pro and ChatGPT. Ranger likens LLMs to historical sophists, useful for solving specific problems but not for seeking wisdom or truth. He emphasizes that LLMs are valuable tools for tasks where output can be verified, speed is crucial, and the stakes are low, provided users remain mindful of their limitations. The article also touches upon how LLMs can reflect the biases and interests of their creators, citing examples from Deepseek and Grok. Ranger cautions against blindly trusting LLMs, especially in sensitive areas like emotional support, where their lack of genuine emotion can be detrimental. He highlights the potential for sycophantic behavior in LLMs, which, while potentially increasing user retention, can negatively impact mental health. Ultimately, the article advises users to engage with LLMs critically, understand their underlying mechanisms, and ensure the technology serves their best interests rather than those of its developers.

Link: https://kagi.com/summarizer/?target_language=&summary=summar...

>> basically all the money will have been spent on Nvidia GPUs that depreciate to 0 over 4 years

I agree the depreciation schedule always seems like a real risk to the whole financial assumptions these companies/investors make, but a question I've wondered: - Will there be an unexpected opportunity when all these "useless" GPUs are put out to pasture? It just seems like saying a factory will be useless because nobody wants to buy an IBM mainframe, but an innovative company can repurpose a non-zero part of that infrastructure for another use case.

I'm not trying to be annoying, but surely if you'd justify spending $200/developer/month, you could afford $250/month...

The reason I wonder about that is because that also seems to be the dynamic with all these deals and valuations. Surely if OpenAI would pay $30 billion on data centers, they could pay $40 billion, right? I'm not exactly sure where the price escalations actually top out.

Safe to assume those are your kids?

Was curious if a maple reserve existed. looks like Canada is leading the charge on that one: https://ppaq.ca/en/sale-purchase-maple-syrup/worlds-only-res...

Same, except I wear this: https://www.amazon.com/Timex-T49950-Expedition-Shock-Vibrati...

It's a "dumb" Timex watch, but also vibrates. So you get that same nice gentle vibrating without any "smart" alerts.

that statement really bothered me. they can of course say that they don't see any evidence of exploitation, but this kind of personal data is valuable to bad actors because they can take it from au10tix and then use it to exploit other services or the individuals directly. au10tix would never know about that exploitation.

Anytime I find myself being clever, I'm reminded of this deadpan exchange.

Tyler Durden: How’s that working out for you? Narrator: What? Tyler Durden: Being clever. Narrator: … Great. Tyler Durden: Keep it up, then.

I was curious about the aggregators. the ones I found referenced in the findings: https://zumigo.com/ https://www.locationsmart.com/ and https://www.microbilt.com/

Anyone using these vendors noticed any weaker data signals/availability that could be related to this? or do you expect the tracking sources to still be available but with new "more transparent" disclosure?

This is reassuring. I never thought I'd be able to think or solve problems, but this gives me hope!

Where do you find/engage with similar high-effort content now (if anywhere)?

>>more is better, right?

yes. always.

"matter of a few months" / "on par offerings"... My (very very) limited impression is both of those statements aren't fair. Google/intel/etc have been working on lots of AI-related projects, even LLM efforts over years. It's not like any of these companies are bootstrapping LLMs or just posted their first AI engineer job a few months ago.

I also don't think I'd call what they're offering "on par" with GPT-4. They're competitive, or impressive substitutes, but my personal tests of Bard at least aren't getting me as useful results as GPT-4.

Source: too lazy to look up any. this is just my lame impression. feel free to downvote.

Agreed. I used to have a diet eating the rich, but after I found out about the greenhouse gas emissions from needed to produce just one free-range rich person, I've switched to ramen. /s

Screw discipline. what you need is my paid medium subscription "the power of YES".

Currently accepting new subscribers at www.notarealnewsletter.com/couldbearealnewsletter/subscribe?utm_medium=email&utm_source=yc&utm_content=1249871092

Just going to state the obvious: this wouldn't have happened if they were walking to starbucks. blatant classism.

The Apple Watch has to have an iPhone paired to support calls which means you need to own and activate an iPhone with a carrier.

I tried to get this setup with a cellular Apple Watch and a cellular iPad but it ultimately meant I didn't have a cellular phone number that could enable the Watch to take calls.

Could protect the moth population with a more liberal policy to provide bears access to picnic baskets...

Also, did you want us to login / create an account? I show up and get a timer, but couldn't really tell what was supposed to happen other than immediately signing up for an account.

I hope this isn't an acquihire. Would be a shame if they shut down the k-cup business.

Part of this looks interesting - is this making it easier for me to organize my browsing? I've been trying to keep track of links for a work project as I browse, would this make it easier for me to store those sites to share or find later? That could be hugely useful.

Or does this make it easier to share those links with other people? If it's just for sharing, I can't tell if it's like Shareaholic or something new.

I'm having a hard time figuring out if this is more for publicly sharing content/links or if it's for my personal storage and organizing? Looks interesting enough for me to look at the beta - send it my way if you want.