Gareth321

I'm with the other guy. Humans exploring space is fun an exciting to me. Sending drones into space is cool, but far less fun and exciting. I want to explore space. I want recreational spacecraft for humans. I want technologies which allow humans to explore the outer regions of our solar system. You might think flying to Mars is a waste of time but I think that's incredible. I would have signed up before I had a family, even if it were a one way trip.

I think it is widely accepted that power corrupts. It is the basis of democracy, and why we would defend it so jealously. Democracy was fought for and won, and we should never forget the alternatives are much worse. This requires peaceful debate, free speech, and free and fair elections.

I agree. This seems like another manufactured internet culture war. One of the most interest podcasts I listened to was around his ideas on democracy and aging demographics. Specifically, as many nations become older, older voters are voting for conditions which benefit themselves at the expense of the young. As they make conditions for the young more and more intolerable, they have fewer children and exacerbate the demographic issues. The effect is self-reinforcing and inexorable. He posits, "maybe this is just how democracies end?" Many Western nations have been using high immigration to plug those demographic gaps, but this isn't a permanent solution. Nations are facing ratios of two working people for every retired person in the coming years and decades, and unless we see some pretty massive technological improvements (or radical economic transformation), things are going to collapse.

I also very much appreciate his positions on how we've allowed physical engineering and sciences to stagnate. It's true. We flew to the moon 57 years ago and basically gave up exploring space. As a space nerd, SpaceX is a breath of fresh air.

Most publicly listed companies in the West require independent audits. Those audits are published for everyone to read. The listed company and the auditors are all tightly regulated, and while fraud can happen, it's relatively rare.

I think you're making the mistake of thinking that because fraud occasionally happens in the West, it's on anywhere close to the same scale as China. If Chinese companies were capable of passing independent audits, they'd all be publishing independent audits. That would be a massive confidence booster in China and internationally, and stock prices would rise significantly. They don't, because they wouldn't pass.

I agree. I think the future for privacy conscious people is going to be an adversarial open weights model, where competing models (ideally from opposing superpowers and labs) critique each other's work to find consensus. Or at least alert for suspicious behaviour.

It could, but it's an inexact science. A little bit like tampering with the brain. Even a surgical procedure removing a tiny tumor can interfere with critical pathways. Usually abliterated models end up almost exactly as capable in benchmarks, but it requires the person doing the abliteration to anticipate or know about the censorship. In the case of Chinese model censorship it is believed that it is post-training, so the underlying data input is not the issue. It's how the model has been trained to use that data. Specifically, abliteration usually targets the refusal behaviour. I would guess that this example could be significantly improved, but I have not seen any abliteration attempts re this specifically.

I partly agree. In support of your scepticism, I also think the frontier labs are overvalued. The expensive part which is not currently priced into the subscriptions is the training. Inference is usually profitable now - at least from what we can tell from industry reports.

Where this bubble differs from the dot-com bubble is that valuations have very strong, organic user demand behind them. In 1999, prototype websites with few customers were getting IPOs worth hundreds of millions. This time, demand is insatiable. Frontier mdoels also has very low global penetration, relatively speaking, so the TAM permits exponentially more sales, in theory.

However the risks are numerous for these labs. Foremost is that Chinese labs are able to easily distill frontier models. A tiny fraction of the cost of training, and they have a close approximation. Second, models are getting much more efficient per task. Companies are already discovering what "good enough" looks like, and it often ends up being much cheaper than what Anthropic and OpenAI are selling. This will only accelerate.

My investment thesis is bullish on the entire AI stack except the frontier labs. I think power generation is going to see huge demand. Ditto for infrastructure and data centers. Ditto for the products and services created by these models.

I would love to hear more about your setup. Which hardware and models? How much context can it handle? How does it compare to frontier models?

Thanks for the tip! I've already got a home server set up. Maybe this is my next AI purchase. Still, US$2,600 here in Europe is not cheap :-/

> No political censorhip? Try asking Chinese models about Taiwan.

Self-hosting means being able to use abliterated models where the superficial instructions like not describing Taiwan as independent are burned off.

Censorship and bias are related but not synonymous. I agree that it's harder to remove political bias when the training data includes it. We know many of the frontier models are heavily trained on Reddit, for example, [which likely explains why we see such heavy left wing bias.](https://trakkr.ai/bias/findings)

More citations:

https://www.gsb.stanford.edu/faculty-research/working-papers... (article: https://news.stanford.edu/stories/2025/05/ai-models-llms-cha...)

https://www.tandfonline.com/doi/full/10.1080/29974100.2026.2...

https://ieeexplore.ieee.org/document/10817610

https://www.washingtonpost.com/technology/interactive/2026/0...

https://link.springer.com/article/10.1007/s11127-023-01097-2

It will not do that unless you specifically teach it to do that, or you use an Anthropic model. The issue in this case was Anthropic's safety monitoring built into the model. They record everything and flag potential illegal activity for human review. This occurs even over the API. I do not know if OpenAI does this, but I would not be surprised. I think the Chinese models are far less likely to do this (unless one resides in China).

It's a good first step. They're not great but they kind of work for many tasks. I think the small models can be useful in context-specific tasks but the friction is high right now because the user is forced to determine whether Siri is good enough, or a better LLM is required per task. Ideally Siri would be smart enough to intelligently route. Apple claims that's how it works but even their cloud processing is far behind frontier models.

Worryingly, they're taking a walled garden approach, and they don't want to expose user data to competitors. After the monumental failure of Siri over more than a decade, I don't trust Apple to execute competently on AI.

The inference is profitable. It’s the training which is really expensive and not being reflected in the subscription prices.

Also no political censorship, and it doesn’t report you to the police when it suspects you did something illegal. That really happened, by the way.

Now I want a local LLM server :( $25k is just so much money.

The last time I saw research on this, they found that the median age of Redditors had dropped significantly over time. I suspect they have a newer and younger user base now. I still talk with older users from time to time and we believe our views on free speech and censorship have not changed significantly over time. However Reddit policy and moderation and majority opinion have changed a lot. They are far more accepting and willing to use censorship to realise political goals.

I’m well aware that Reddit doesn’t reflect reality. Most of society is not far left, chronically online, violent, and out of touch with reality.

The study you provide isn’t related to my comment but if you’d like to shoehorn it into future discussions you should at least read it first. That paper is widely criticised, and is, ironically, the kind of thing shared on Reddit. I’ll just copy and paste existing academic criticisms of the paper because I don’t really think you’re going to take the time to seek this out yourself.

1. Inconsistent data collection: The far-right data came from an established extremism database, while the far-left data had to be assembled separately because the database did not routinely track most far-left extremism. This creates a potential sampling bias, with no demonstrated guarantee that both sides received equally comprehensive coverage.

2. Serious numerical inconsistencies: The study counts 168 deaths from the Oklahoma City bombing but reports zero far-right bombing incidents. It also contains basic arithmetic errors, including percentages totalling 110%. These inconsistencies raise legitimate concerns about data quality and the reliability of the analysis.

3. Subjective ideological classifications: The study uses broad political definitions and includes homicides with weak evidence of ideological motivation. Without a fully accessible case-level audit trail, independent researchers cannot easily verify whether incidents were classified consistently or accurately.

4. Unsupported statistical conclusions: The author suggests far-left attacks are more likely to cause multiple fatalities, but the reported difference is not statistically significant (Fisher's exact test, p ≈ 0.22). Drawing substantive conclusions from this difference is therefore questionable.

5. Disproportionate influence of one attack: The Oklahoma City bombing accounts for approximately 32% of all far-right fatalities in the study. Removing that single incident reduces the far-right-to-far-left fatality ratio from 6.7:1 to 4.6:1. The overall disparity remains, but its magnitude is significantly influenced by one exceptional event.

6. Selective historical timeframe: The study begins in 1990, excluding the 1960s–1980s, when far-left revolutionary movements were especially prominent in American domestic terrorism. This timeframe may be appropriate for the research question, but it makes broad claims about political violence across modern history misleading.

7. Conclusions beyond the evidence: The study examines ideologically motivated homicides, not political violence generally. It excludes non-fatal assaults, riots, intimidation, arson without fatalities and other violent offences. Using its findings to claim that the far-right is responsible for most actual violence is therefore a substantial and unjustified extrapolation.

Notably, the time frame used doesn’t reflect reality today, so it cannot be used to make normative claims about which ideology is more violent right now.

To be clear I’m not suggesting that that is my preferred outcome. I’m worried that that is what this new generation will vote for. Then when they get it, they’ll realise what they lost, and they’ll have to fight to get it back again.

No, and it’s very rare for Chinese companies to allow independent audits. I would not trust DeepSeek financial disclosures. However from industry data we can see that inference is often profitable. It’s the training which is very expensive. Which is not a big problem for Chinese companies since to date much of their training is just distilling frontier models.

As a now former Chinese investor, the numbers reported in China are almost entirely fictional. They are whatever the local prefect or senior leaders wants them to be. Right now they seek to imply to the West that they are competitive, so the numbers reported reflect that. Almost no Chinese company permits independent audit of any of their finances.

It’s a rabbit hole but if you have some time it’s fascinating to look into the sentiment of investing in China vs the rest of the world. Because the numbers are so unreliable, investing is considered about the same as gambling in China. You’re more likely to lose a lot of money than win, and there is no way to accurately predict strong from weak companies based on their reported financials. Occasionally senior leaders will sentence the leader of very large failed company to death to attempt to maintain confidence, but it’s not working. This is one of the reasons for the housing bubble. There just aren’t stable investing opportunities in China.

They published all/most of their work and it would be unbelievable if Anthropic had not already used that research in their own architecture.

The flattening part is not reported, but based on estimates from a company called TickerTrends. If they were correct it would imply a sudden and unusual departure from some incredibly impressive growth. Even more unusual given the release of their industry leading Opus/Sonnet/Haiku. Something to watch but I would not consider that authoritative, and would not invest based on that either.

Price per task has been dropping exponentially for years. This isn’t a sudden new phenomenon. They’ve been competing with each other continuously. There is even more competition now than ever before.

I agree. Interesting you’re being downvoted. I’ve been using Reddit for 16+ years and have watched it morph from a libertarian, centrist community dominated by developers and geeks, into a very left wing, violent, vehicle to export American political activism to the world. Redditors used to abhor censorship. People would practise debating all kinds of taboo topics because this is how we become wiser. Now just commenting on the wrong community will see you automatically banned from 30 of the largest subreddits.

They study this. [Gen Z is 10x more accepting of violence to suppress speech they object to compared to Boomers.](https://expression.fire.org/p/gen-z-is-10-times-more-accepti...) It seems history really does repeat. People are going to need to experience some authoritarianism to remember why we care so much about free speech.

It’s hard to take down sites on Tor. Usually impossible without the full weight of the FBI/CIA. I agree with the user above: we can do decentralised communication without censorship. Of course if you start from the faulty premise that online discourse needs to occur on one of a handful of censored platforms then that’s impossible.

> In the macroeconomy, who is the "efficiency" we are chasing for, who benefits? Who is getting the money that was "saved"?

Everyone who uses those products and services. Prices go down and people can get much more in aggregate. More entertainment, healthcare, etc.

My media server is one of my favourite things in the whole world. Endless content in high bitrate video and audio. I have the editions I like (director's cut of Kingdom of Heaven is INCREDIBLE). No ads, ever. I watch it wherever and whenever I like. It works on all devices, online and offline. I'm never geofenced. Content will never be removed for any and no reason. Best of all, it's all in one place: movies, shows, and music. I can't pay for a service like that - and I probably would have, if I could.

I was paying for Netflix, but through a thousand cuts they pushed me away from streaming platforms entirely. I probably spend more on hard drives now than I would have with streaming subscriptions, but it's an infinitely better service.

My two favourite apps are Sonarr and Radarr. They make the collection process seamless.

> China emits half the emissions per capita of the US.

China emits more CO2 than any other country by a wide margin. It's not less damaging to the environment just because they have more people. Even if we pretended that that were the case, [China emits more CO2 per capita than 32 of 38 OECD countries.](https://ourworldindata.org/grapher/co-emissions-per-capita?c...) They are an egregious and unrepentant polluter, in every way imaginable, and in many ways you cannot even imagine.

Any discussion of reducing CO2 emissions which doesn't include a large section on China is not taking the subject seriously. But to be clear, we should also discuss emissions by the other large polluters like the U.S. and Canada.

> The rate at which we're churning out AI-generated code is going to make this worse, too, because AI-generated code is kind of notoriously terrible at security.

I think this was true in the past but AI is helping us finding exploits which have been in our code for many years. All out there by humans. We’re seeing fewer customer reported issues, and despite the pen tests getting far more rigorous, we’re passing all of them. Obviously this is one data point, but I’m hearing similar things from my friends in security and ops.

I'm not worried at all. LLMs are already so good that they can reproduce many apps and games merely by using those apps and games for a little while. By the time customers are willing to accept streaming only games (which I contend is still MANY years away from mainstream), LLMs will be able to near perfectly replicate these games. Or better yet, make far better versions of them. I'm not even sure we have a games industry in a few years.

It sounds French.

I just check. It is in fact French. Well, the company is French.