techpineapple

I really find tech "optimism" distasteful. It's become this weird battle ground where I dunno, if you're reasonably optimistic but think AI isn't set to become a god you're called a doomer, at the same time that Tech optimists are basically giving up on humanity.

I guess on a more interesting side - it feels interesting more and more how our polarization leads us to see our side as inevitable. It seems inevitable to me that there will not be a fast takeoff, that's just the way that the universe works, that things are incremental, that things grow fast for a while and then taper off. And as someone who believed this, it still is super _weird_ seeing all the people say no, AI is different, it's going to lead to Utopia.

As weird as it is for people who read this and believe the opposite think I am. But I think this is also a reaction to where we are as a species. Everyone wants the easy answer.

On some weird level, it's weird to me that the same psychological quirks that inhabit folks and make them believe in sort of cults or spiritual beliefs also inhabit billionaires, and make markets spend billions of dollars.

But yeah. This article is... Real is beautiful! The possible is beautiful! AI is cool, let's enjoy it and all the other cool things in life.

Interesting I came to the exact opposite conclusion. In addition to reducing sycophancy they reduced character, and the margins on consumer products will never be enough to justify the levels of investment.

I’ve heard of this idea of training on synthetic data, I wonder what is that data and does this increase or decrease hallucinations? Is the goal of training on synthetic data to better wear certain paths, or to increase the amount of knowledge / types of data.

Because the second seems vaguely impossible to do.

I wonder if one reasons new versions of GPT appear to get better - say at coding tasks is just because they have new knowledge.

When ChatGPT4 comes out, new versions of API’s will have less blog post / examples / documentation in their training data. So ChatGPT 5 comes out and seems to solve all the problems that ChatGPT4 had, but then of course fail on newer libraries. Rinse and repeat

Let’s say we’re in some kind of weird new tech/AI bubble or whatever. Obviously substack gains from being invested in, even if the bubble pops. Do the investors gain?

Wouldn’t this dramatically limit the governments power to influence, well everything? How much green energy investment went to companies with patents? Tesla and SpaceX would be out.

I feel like for the broad median of us, there’s a balance between giving a shit and healthy detachment. Unless you’re at the top, I don’t think working 90 hours a week in with tunnel vision will make you happy, you’ll probably end up burned out and possibly cast aside in a reorg.

On the flip side, I cringe at the posts that go too far in the other direction emphasizing it’s just a job and the job doesn’t love you.

It’s true, on its surface it’s good to have healthy detachment, but it’s also good to take risks, make yourself vulnerable and give a shit. Yes, know that it’s a transactional relationship but then give a shit within that framework. Give it your all, make friends with your coworker, be moderately devoted to the company in the moment, while respecting your own definition of work-life balance and being ready to say goodbye when it’s over.

Seems like maybe wang didn’t give a shit.

I’ve noticed inconsistencies like this, everyone said that it couldn’t count the b’s in blueberry, but it worked for me the first time, so I thought it was haters but played with a few other variations and got flaws. (Famously, it didn’t get r’s in strawberry).

I guess we know it’s non-deterministic but there must be some pretty basic randomizations in there somewhere, maybe around tuning its creativity?

Noticed that Most of this Reddit AMA is about how great 4o is and how terrible 5 is in comparison:

https://www.reddit.com/r/ChatGPT/comments/1mkae1l/gpt5_ama_w...

I suspect this is a symptom of something bigger and not a cause. Our hyper-capitalist world has become so transactional — of course when your value and ability to participate in this world are based on your ability to have a certain set of character traits, not having those characteristics will be seen as a “problem”

I have ADHD. I take meds so that I can do my job and get paid to afford to participate job society. I don’t take it on the weekends, and yes, I have a personality that my family loves, but the rest of the week? It’s not cute to be forgetful at work.

And then you have Mike Johnson talking about lazy 24 year olds in their basement playing video games.

We literally can’t afford to be an almagam of traits of our ancestors, when the social safety net going away, entry level jobs being more competitive and the decline of highly paid unskilled labor.

I'm curious if this might help Cursor's lighting money on fire problem?

https://pivot-to-ai.com/2025/07/09/cursor-tries-setting-less...

is this enough of a price difference to make cursor profitable?

It doesn't feel like the country is currently headed in this direction.

Maybe I'm reading this wrong, but why would you choose doordash over a lower paying project management job and work your way back up, especially if they've "fallen behind on their mortgage payments"

Sadly, and as someone who is not a fan of organized religion, I think the decline of organized religion in society has had some weird negative repercussions with everyone looking for new “religion”

Sounds like a money making strategy. Also, given how expensive all this shit is if inference costs _more_? That’s not cheap to me.

But again the original argument was that they can run forever because inference is cheap, not cheap enough if you’re losing money on it.

Did they really have another choice? if no big leap was on the horizon are they just never going to release 5? I mean, from a marketing perspective.

I’m curious what this means. Maybe I’m stupid but I read through the sample gpt-4 vs got-5 and I largely couldn’t tell the difference and sometimes preferred the gpt-4 answer. But like what are the average 800 million people using this for that the average 800 million user will be able to see a difference?

Maybe I’m a far below average user? But I can’t tell the difference between models in causal use.

Unless you’re talking performance, apparently gpt-5 is much faster.

Wait why are these companies losing money on every query of inference is cheap.

After talking to ChatGPT for nearly five hours, and inventing a brand-new physics framework dubbed “The Orion Equation,”

It’s wild that Travis Kalanick was bragging to the all-in folks recently about having a similar experience.

It’s interesting there was a report out a year or two ago that Ukraine was either promoting pride parades or like teaching tolerance for gay folks, because they needed more folks for the military and Ukrainians were homophobic.

Seems like when the military is already having trouble recruiting folks, maybe becoming less tolerant isn’t the solution. Decisions like this make Americans less safe.

This may be a hugely unpopular opinion, but I don’t think it should be legal to sell a product at a loss, except maybe in clearance situations. Like you should be able to sell a product for $100 then use $1 million investment to develop a new process, then sell for $80 with the cheaper process, but not sell for $80 while the cheaper process is being developed.

It’s a stark number: 212,000. That’s how many women ages 20 and over have left the workforce since January, according to the most recent jobs numbers released Aug. 1 by the Bureau of Labor Statistics. (By contrast, 44,000 men have entered the workforce since January.)

  The use of “entered” and “left” is confusing to me, are they comparing apples and oranges or net job gains and losses.  Presumably some women entered the workforce during this time and some men left right?

I wonder if this is to shift the gpus to X.AI?

One weird question is why is it that they can win the math Olympiad but can’t count the r’s in strawberry. I mean I know they fixed the r’s problem, but it seems like they’re really focused much more on raising the ceiling than the floor. In many ways, counting r’s in strawberry and the like are the _more_ interesting problem, and yet that’s not a substantial part of the sales pitch.

Interesting readign the progress.openai.com sample prompts https://progress.openai.com/?prompt=6

I would say GPT-5 reads more scientific and structured, but GPT-4 more human and even useful. For the prompt:

Is uncooked meat actually unsafe to eat? How likely is someone to get food poisoning if the meat isn’t cooked?

GPT-4 makes the assumption you might want to know safe food temperatures, and GPT-5 doesn't. Really hard to say which is "better", but GPT-4 seems more useful to every day folks, but maybe GPT-5 for the scientific community?

Then interesting that on ChatGPT vibe check website "Dan's Mom" is the only one who says it's a game changer.

404?

“Even if AI progress completely stalls today and we don’t reach AGI… the current systems are already capable of automating all white-collar jobs within the next five years.”

I’m not sure if this is playing word games, but this is technically true right? Some aspect of all white-collar jobs can be automated. Just not the soup-to-nuts job.

Also it’s interesting, what is Dario Amodei suggesting politicians do? UBI? A vast expansion of the welfare state? 1 child policy?

“Cancer is cured, the economy grows at 10% a year, the budget is balanced — and 20% of people don't have jobs."

Why only 20% unemployment?

> No serious state would ignore potential conflicts of interest with their biggest adversary related to advanced technology and military affairs:

I agree with you, but then we're not a serious state since we've been doing this near constantly recently. Apply it equally or don't apply it at all.

I’m curious if this problem can be solved. I suppose on the one hand, if some universities can prove that a harder curriculum leads to better outcomes, then yes, but our entire culture is moving away from say inherent respect for the institution to a customer-orientation.

And given what I think was something of a debacle in the entire concept of code schools, can you blame people?

But do I want to pay, I dunno, $60k a year for a degree that might fail me out and even if I succeed is only tangentially related to a job?

In my older and wiser years I might pay a few thousand dollars to take a literature course, but a funny thing about credentials is…. I would take a literature course because of the hope that it might provide some inherent value that would make me a better leader/thinker/etc. not to get me a job. And the credential in some weird way is almost an anti-pattern. Really (very ideally) the measure of a university program shouldn’t be the ability to get a job, but the ability to thrive once there, particularly for the liberal arts.

It makes me wonder, I love watching movies about conservative educational institutions. But I love watching them because they're a fantasy. I wonder if they were always a fantasy. Were universities better? Or was there just less competition. Will a harder english curriculum produce higher quality graduates? Or just filter out the lower quality graduates.