Gareth321

You're going to get a wide range of responses on this but given the improvement in these models in just one year, and the number of bugs they're detecting which humans could not, I suspect that even median vibe-coded software is going to surpass median human coded software soon - if it hasn't already.

The important thing to remember here is there perfect isn't on the table. The benchmark is existing human-introduced bugs vs LLM-introduced bugs. Many developers have encountered odd bugs which a human would not have introduced, while forgetting about all the bugs caught which humans introduced. Or their opinion is formed by models from six months ago.

I am also pretty disillusioned with government right now, and I share your views on common law marriage and alimony, but I see this as one of the few things a government is doing right.

I don't think this is accurate, but perhaps it matters how you slice the data. Women's mental health has been steadily declining across all strata for decades - faster and worse than men's. It's at record lows now, irrespective of whether they are single, married, or with kids. For example, [the UK recorded a 46% increase in mental health issues among women aged 16-64 since 1993.](https://digital.nhs.uk/data-and-information/publications/sta...) Further, when we look at the data between single and married women, [married women in the U.S. General Social Survey (GSS) are about 78% more likely to describe themselves as “very happy”, while never-married women are about twice as likely to say they are “not too happy”.](https://statmodeling.stat.columbia.edu/2024/08/27/marriage-h...)

The irony with all this data is that as women have achieved parity with men in all kinds of domains, their happiness has commensurately declined. This is one of those hot button topics which most sociologists refuse to research.

I think this is the end of sites where it is expected that only humans may interact with them. It's been cat and mouse for a while, and some places like Reddit sell API access, but these agents are for all intents and purposes, humans interacting with the site. They're going to need to figure out new business models.

Most residential proxies are already far more blocked and rate limited than any Meta IP. The internet is becoming a very weird place, where individual and "trusted" personal IPs are becoming a kind of commodity. Some sites are already scoring IPs based on usage activity - like a credit score. It's only a matter of time until this data is collated and commoditised. AI analysis is turning this up to 11.

Honestly, the levels of desperation and despair I'm seeing in the hellscape that is modern dating, I think a lot of people would sign up for this. I probably would if I were single. It's very similar to how matchmakers work, and they're popular all over the world, and have been for millennia. They're an unbiased and impartial judge and coach. We all have enormous biases and weaknesses in so many ways, and we could all do with help to better ourselves.

The important distinction here is no one is forced to marry anyone. This is entirely optional.

Israel is a notable exception. They're well above replacement, even if we remove the Haredi. Women have full equality. I'm not sure what that means though. That healthy fertility requires constant war? Or that it requires high religiosity?

You continue to soapbox without addressing the content of this discussion. I'll try one more time.

> Meetoo did not meaningfully changed peoples ability to "form relationships" in work.

You are confidently incorrect. [People widely reported a stifling effect on office relationships.](https://legacy.vault.com/blogs/workplace-issues/2018-vault-o...) More than 1 in 4 said #metoo made them less accepting of workplace romance. 43% considered relationships between workers at different levels unacceptable - 15% higher after #metoo.

> Meetoo did not even meaningfully changed ability to abuse, abusers even had their own backslash.

This has nothing to do with what we are discussing. I understand you have strong feelings about this but I need you to understand that from the bottom of my heart, I do not care about your feelings.

> Pre-metoo, while people sometimes formed relationships at workplace, it was not the primary way to form relationships. Plus, companies did not wanted the drama from predictable breakups anyway (which is the primary source of issues with workplace dating).

No one claimed it was the primary mode of forming relationships.

> This is triple funny on HN, where most men work in male dominated or male only teams.

This is one step removed from, "men are stupid, am I right, ladies?" Your disdain for men, the users on Hacker News, or whatever villain you've concocted in your mind prevents you from having a normal conversation like a normal person.

This has been a thoroughly uninteresting waste of time. I suspect you became lost on your way to some protest or /r/Politics.

Perhaps we should blame this on shitty website design. So many of them load huge assets for no good reason. They get away with it because modern processors are that good, bandwidth is cheap, and until now, RAM has also been cheap.

I find this oddly comforting :)

This might be how it shakes out but I don't think it's acceptable, and I don't think white collar workers will think it's acceptable. Taking high income people and making them low income people (usually for life) has a tendency to politically radicalise people. Especially in light of the unprecedented productivity and profits which businesses will experience. People will demand a slice of the pie.

My naive hope is democracy, so we need to defend it.

Yes, and increasingly, data will no longer be a moat either. These models will be capable of creating all the data/content we could ever want. There will always be a market for "organic" movies with real actors, but we should expect the value of that to decline precipitously.

My investment thesis is currently around these strata:

1. Electricity. All analysis shows massive shortages beginning now with no end in sight. Western governments in particular have massively dropped the ball on this one.

2. Chip design and production. Nvidia and TSMC dominate here for a few reasons. Partners required in the supply chain like ASML benefit, but they're not really the bottleneck right now. We're seeing a lot of investment pouring into this space and we'll see competent design competition (especially from Google's TPUs, and potentially from Apple and AMD). Chip fab is very difficult to get right, and takes a long time to ramp up. See the NAND space.

3. AI labs. These guys are doing to get commoditised. They're trying very hard to prevent distillation but I think it's impossible. The open weight models are not far behind, and the Chinese labs are getting very creative about how they work with limited hardware. If anything, they'll decrease inference cost over time, reducing the value of existing data centers (of which OpenAI and Anthropic have purchased an enormous amount). I am very bearish on the future value of OpenAI and Anthropic.

4. Software. So much software. Everyone gets software. The companies which benefit from this will be able to monetise it. Google, for example, might put ads into VERY capable and free AI. People might accept that. Microsoft has been investing heavily into their data centers, and maybe that's their revenue center in the future. Especially securely hosted services. I expect massive growth in this space in the short term. As long as people still have high paying jobs, they'll pay for well made software and tailored services.

I don't think you read my comment before you replied (or the comment I replied to). I encourage you to take another look. It's hard to meaningfully respond to a non-sequitur.

To reiterate, the social cost discussed is the inability for people to form relationships at work. I acknowledge all of the costs of people being harassed at work. That's bad. I hope you read this comment and understand that I am not disagreeing with you.

This is easily refuted. [In 1950, food, transport, and housing consumed 70.3% of median working household expenditure. Today is it 63.3%.](https://www.bls.gov/opub/100-years-of-u-s-consumer-spending....) In 1950, 87% of family households were married couples, and only 21% of married women participated in the labour force. Today, 62% of cohabiting couples have both partners working. Households are earning more, but prices rose almost commensurately. If you were correct, the proportion spent on necessities would be far lower.

I will concede that luxury goods are far cheaper, and perhaps one could make an argument that cheap TVs are a reasonable concession for expensive food and housing. I would not agree with that.

If I understand your question correctly, it reduces the value of software to almost zero. Meaning all of the SaaS companies are going to have to work very hard to diversify their revenue streams into value added services. That's going to be very difficult or impossible for most to navigate. Hence the term in the industry "SaaSpocalypse."

Some companies will retain a moat. For example, YouTube holds incredible IP, and until AI can replicate that wealth of content at scale for cheap, it will remain profitable. Facebook has the user critical mass. Increasingly, hardware for inference is going to be a moat. So is electricity.

Which loops back to my comment above: this breaks a lot of the market dynamics required for a functioning economy. Accelerationists are convinced that we'll invent new jobs to replace the old jobs, but I don't think that's going to happen. Certainly not at the pace of change we're seeing.

That sounds fun but this is people paying for real products and services. The demand is there (and it's growing faster than any other commercial technology in history). They just need to work on inference (and potentially training) costs. I think that's clearly achieveable.

For the record, growth plays have been common in Silicon Valley for many decades. It's how we ended up with all the major tech companies we see today. Of course, it's high risk, high reward.

I am in two opposed minds on this. On the one hand, you're correct. Their job will be made redundant in a couple of years. Some senior people might be retained to orchestrate all the agents, but eventually they will also be replaced. And it's coming for all white collar jobs, not just coders. That's bad.

On the other hand, unlimited intelligence democratises products and services in a revolutionary way. Everyone will be able to do whatever they like in digital space for basically free. No more purchasing software. Make any app/company/website/blog/movie/show/game/forum you can think of. This will quickly bleed into the physical space, with real-world products and services becoming logarithmically cheaper. A true future of abundance.

The big problem is going to be our social and economic transition. Are we going to wait until unemployment is that 10% before we act? 20%? How bad are we going to allow things to become before we implement universal high income? I say universal high income because white collar workers are not going to accept being relegated to an income of $1000/month forever. To achieve universal high income requires eye-watering taxes on the major benefactors of our new economy. They won't accept that without a fight, and it our current fractured political climate, what would it take for everyone to agree on a unified direction?

Google stands to lose enormous revenue from people not using search - and we already see traffic declines. It's truly an existential threat for them. Their competencies are in advertising, and if people are going to be using chatbots in future, they're going to have to advertise in their chatbot. That sounds awful to me, but maybe people will be cool with it?

I've had many similar experiences. It's confidently incorrect to a shocking degree. Worse than ChatGPT from two years ago.

I agree. It's much worse than the cheap Chinese models. They appear to heavily bias parametric knowledge and discourage tool use. That's fine for things like "how do I perform CPR?" but worse than useless for any kind of research. [There is one benchmark showing far lower rates of hallucination, so let's see how accurate this is.](https://www.reddit.com/r/singularity/comments/1wuj72j/gemini...)

While that's annoying, the other frontier models easily overcome this with appropriate tool usage. I do a lot of research with frontier models and they're very good about identifying where their parametric knowledge is insufficient and searching for the correct knowledge on the internet. 3.8 Flash is HORRIFIC. The majority of the time it doesn't use any tools and infers things from its parametric knowledge. Things which should clearly have implied tool calls. Historical statistics, legal precedent, economic data, etc. I think it's incredibly clear that it has been tuned for speed and not accuracy.

Of course, it's called "flash," and that implies its purpose. I have little use for speed and a LOT of use for accuracy, so I'm hopeful 4.0 is much better. I saw a benchmark earlier today showing that it is much less prone to hallucinations. Let's see.

I have met very few men for whom prospective partners' body counts is a red line. It might be a preference, but it's always in context of the larger picture.

I think there has been much topical research on this now, and from what I can glean, there are at least two forces working at the same time.

1. Gendered preferences didn't change, but society did. Women still prefer men who earn more than them (by a very wide margin). This would be okay if we still lived in the 1940s, but childless women outearn men now in most metro cities. Sometimes by a wide margin. Women are graduating in high-paying professions at ratios of 60:40, 70:30, and even higher, meaning this problem is getting much worse. Increasingly, the majority of women are focused on the top 10-20% of men who outearn them, to the exclusion of all other men. This is compounded because incomes are just one of many selectors. There is also height (which remains a surprisingly strong preference for women), attractiveness, many personality traits, social standing, wealth, political affiliation, etc.

2. Dating apps have made the above exponentially worse. It's very easy for a woman of average attractiveness (I say this in a academic sense, not derogatorily) to only swipe right on the most attractive men. Those men are perfectly happy to have a lot of casual sex with said women, but would never consider them for marriage. This has warped the perception of their "league." Many of these women end up 40 and alone and deeply unhappy. Or they "settle" for a man of average attractiveness and income and resent him.

This is almost impossible to untangle. We don't want women to be less educated, or earn less. We don't want them to feel like they're "settling." At the same time, there are many amazing men out there in their league, and they could have a happy marriage. I applaud Singapore trying something, because the future of the West looks like a lot of angry, bitter, single, childless, lonely people.

It's just another example of social change which was made for good reasons but has resulted in negative outcomes. Companies wanted to reduce the risk of accusations of workplace misconduct in the wake of #metoo. This reduced misconduct, but at massive social cost.

Another example is that most economists admit that as women entered the workforce en masse, prices settled higher for the same goods and services. Now survival requires two incomes instead of one. It was a massive net loss to society, but we also agree that women should be free to work if they would like, without barriers. We see similar things with women becoming educated. The more educated women become, the fewer children they have. That's really bad for society, but we also want women to be free to become as educated as they like.

Sadly I see no easy solutions here because it requires weighing individual and societ-wide rights. Companies aren't going to risk higher rates of misconduct to benefit society broadly. These are collective action problems, and unless we find ways of individualising the problem, they're only going to get worse.

I will meet you half way. The line between base load and backup capacity is semantic. Most modern power plants (including nuclear) can ramp up and down production depending on demand. At the high elasticity end is gas generation which can ramp down to 0%, and at the low elasticity end are old nuclear reactors which might only be able to modulate 10-40%.

As you infer, it doesn't really matter whether you read my comment as base load or backup capacity, and I agree with the rest of your comment. Though I would have my own nit pick with "exceptional circumstances." Some base/backup generation is required in almost every modern grid almost every day. That's not exceptional. That's normal operation.

I'm not sure how much of a benefit this is. Many/most houses/commercial sites won't have batteries (or at list some significant minority). They'll still complain. Power companies will still have a duty to restore power ASAP for health and safety reasons. Either way, they'll only repair the lines when it's safe to do so, and they'll continue doing that.

Which is deeply ironic, given the content.

I'm the PM side of this debate with my TL. I've tried meeting him on every level of understanding. Graphs, story telling, Jira issues, business cases, ITIL, OLAs and SLAs, customer interviews, user interviews, you name it. He strongly resists having to justify the things he wants to work on. His justifications are usually some combination of "it's too complicated to explain," and "this code is a mess." Problem is, I need to account for their hours. I need to explain why we're spending money on that thing - to people who are even less likely understand what my TL is saying.

I say this not to diminish your frustration but to appreciate that you tried to meet him at his level, and provide some context for what the other side can feel like.

> Cache doesn't help you much when you are compacting every 5 minutes...

It's crazy on Codex. I sometimes get just 2-3 turns before it compacts. It has forced me to use persistent project documentation for everything. Maybe that's not a bad thing but unless it reads all the documentation after every compaction (and uses half its cache), it goes off the rails. By comparison, Opus 5.5 is a breath of fresh air. It takes FAR longer to hit the cache limit and that means it keeps useful information in working memory far longer. I think this alone has resulted in a massive productivity and efficiency increase for me.

I agree on all points. This is a short term cost saving measure at the expense of long term energy security. Batteries do help smooth the considerable volatility introduced by renewable generation, but unless they want a significant share of houses and industrial sites to have large batteries (which would cost FAR more than just keeping the power stations online), they still need base load generation.

Unless this is very carefully managed (and I do not think governments are generally capable enough), this will result in a lot of new problems. One will be major difficulties in developing new housing (or intensification) due to lack of electricity supply. Another will be insufficient supply during 99th percentile events in which home batteries alone will be insufficient to close the gap. Another will be distribution. Some neighbourhoods will have higher battery penetration than others, and this happens organically and without the permission of the grid companies. That's very difficult to plan around. Some areas are going to be secure, and others will not, unless the grid purposefully oversupplies, negating the potential cost savings.

Batteries are good at marginal demand smoothing, but I am very dubious about shutting down energy generation. Especially at this time. All analysis points to the reality that due to technology demands, we're facing an imminent power generation crunch. Prices are going to rise everywhere and demand will only increase.