AJ007

I've been around for a while (not quite retirement age) and this time is the closest to the new feeling I had using the internet and web in the early days. There are simultaneously infinite possibilities but also great uncertainty what pathways will be taken and how things will end up.

There are a lot of parts in the near term to dislike here, especially the consequences for privacy, adtech, energy use. I do have concerns that the greatest pitfalls in the short terms are being ignored while other uncertainties are being exaggerated. (I've been warning on deep learning model use for recommendation engines for years, and only a sliver of people seem to have picked up on that one, for example.)

On the other hand, if good enough models can run locally, humans can end up with a lot more autonomy and choice with their software and operating systems than they have today. The most powerful models might run on supercomputers and just be solving the really big science problems. There is a lot of fantastic software out there that does not improve by throwing infinite resources at it.

Another consideration is while the big tech firms are spending (what will likely approach) hundreds of billions of dollars in a race to "AGI", what matters to those same companies even more than winning is making sure that the winner isn't a winner takes all. In that case, hopefully the outcome looks more like open source.

Right, this is having Claude Code just running as an agent doing a lot of stuff. Also tool use is a big context hog here.

Unfortunately it's called war and it appears to be part of human nature.

Pretty easy to hit $100 an hour using Opus on API credits. The model providers are heavily subsidized, the datacenters appear to be too. If you look at the Coreweave stuff and the private datacenters it starts looking like the telecom bubble. Even Meta is looking to finance datacenter expansion - https://www.reuters.com/business/meta-seeks-29-billion-priva...

The reason they are talking about building new nuclear power plants in the US isn't just for a few training runs, its for inference. At scale the AI tools are going to be extremely expensive.

Also note China produces twice as much electricity as the United States. Software development and agent demand is going to be competitive across industries. You may think, oh I can just use a few hours of this a day and I got a week of work done (happens to me some days), but you are going to end up needing to match what your competitors are doing - not what you got comfortable with. This is the recurring trap of new technology (no capitalism required.)

There is a danger to independent developers becoming reliant on models. $100-$200 is a customer acquisition cost giveaway. The state of the art models probably will end up costing hourly what a human developer costs. There is also the speed and batching part. How willing is the developer to, for example, get 50% off but maybe wait twice as long for the output. Hopefully the good dev models end up only costing $1000-$2000 a month in a year. At least that will be more accessible.

Somewhere in the future these good models will run on device and just cost the price of your hardware. Will it be the AGI models? We will find out.

I wonder how this comment will age, will look back at it in 5 or 10 years.

This entire thing is a convoluted mess that exists just to employ EU bureaucrats.

The solution is pretty simple: Apple is no longer allowed to run an "App Store" or distribute a pre-installed web browser. This part can be split apart as a separate company if they prefer that to shutting it down. When you look at Apple's taxable income, most of it goes to Ireland. It's very bizarre to call a company that makes everything in China and holds all of its IP in the EU a US company. It isn't.

Apple does what the Chinese government tells it to (see Apple in China by Patrick McGee.) If they pulled the EU stuff over there, Apple would cease to have new hardware to sell.

There are other good use cases here like documenting recurring bugs or problems in software/projects.

This is a good illustration of why e2e encryption is more important than its ever been. What were innocuous and boring conversations are now very valuable when combined with phishing and voice cloning.

OpenAI is going to use all of your ChatGPT history to target ads to you, and probably will have to choice to pay for everything. Meta is trying really hard too, and already is applying generative AI extensive for advertiser's creative production.

Ultra targeted advertising where the message is crafted to perfectly fit the viewer mean devices running operating systems incapable of 100% blocking ads should be considered malware. Hopefully local LLMs will be able to do a good job with that.

Yeah, a better title for the article is how western automakers are going to go extinct. Sure the US might decide to block Chinese cars (apparently the EU isn't), but they can't force the rest of the world to buy $65,000 American built cars when the alternatives are less than 1/3rd of that price.

A larger question is how much the cheap Chinese cars are dependent on a long chain of government subsidies from the mines to the local infrastructure and what happens when China's investment driven growth cycle comes to an end. If the solar panels are any comparison, the Chinese automakers are losing a lot of money despite grandiose subsidies.

Growing backlash yet used by hundreds of millions of people and growing.

Very helpful for ad targeting. As Apple kills tracking and ramps up its own ad business, Meta will need to collect as many signals as possible.

I understand what experienced developers don't want random AI contributions from no-knowledge "developers" contributing to a project. In any situation, if a human is review AI code line by line that would tie up humans for years, even ignoring anything legally.

#1 There will be no verifiable way to prove something was AI generated beyond early models.

#2 Software projects that somehow are 100% human developed will not be competitive with AI assisted or written projects. The only room for debate on that is an apocalypse level scenario where humans fail to continue producing semiconductors or electricity.

#3 If a project successfully excludes AI contributions (not clear how other than controlling contributions to a tight group of anti-AI fanatics), it's just going to be cloned, and the clones will leave it in the dust. If the license permits forking then it could be forked too, but cloning and purging any potential legal issues might be preferred.

There still is a path for open source projects. It will be different. There's going to be much, much more software in the future and it's not going to be all junk (although 99% might.)

Contest between Claude Code and Gemini CLI, who rewrites it faster/cheaper/better?

Current best practice for Claude Code is to have heavy lifting done by Gemini Pro 2.5 or o3/o3pro. There are ways to do this pretty seamlessly now because of MCP support (see Repo Prompt as an example.) Sometimes you can also just use Claude but it requires iterations of planning, integration while logging everything, then repeat.

I haven't looked at this Gemini CLI thing yet, but if its open source it seems like any model can be plugged in here?

I can see a pathway where LLMs are commodities. Every big tech company right now both wants their LLM to be the winner and the others to die, but they also really, really would prefer a commodity world to one where a competitor is the winner.

If the future use looks more like CLI agents, I'm not sure how some fancy UI wrapper is going to result in a winner take all. OpenAI is winning right now with user count by pure brand name with ChatGPT, but ChatGPT clearly is an inferior UI for real work.

You can, for now.

Apple has lost the Google deal and the ability to force apps to pay them 30% of sales. Both were ultra high margin. Apple is going to replace that lost revenue by filling the iPhone up with ads.

There's also another point that security is really fucking expensive. Apple on Google spend billions a year on security, yet their phones are broken in to once they are a couple of years old. Big American software companies have large margins and large budgets. Those Chinese companies are running on fumes (and credit.)

Security and encryption is taken as a given by Western regulators given how many times they pass laws to break encryption. If you look at targeted 0-days, the conclusion would be more along the lines of the very best hardware+software is barely secure.

When you start looking at alternative messengers outside of Matrix, XMPP, and IRC, there isn't much where third parties can operate or implement both servers and clients.

Certainly if no one can implement these two things it is functionally a closed source project. It also is a security failure from the standpoint of control, validation, and also future security and vulnerability patching (there's a graveyard of dead "secure" messaging apps.)

Is DeltaChat perfect from a security standpoint? No, but it's certainly well above the hurdle most people are at now. Most people are using non-encrypted communication that is actively scanned & stored, or e2e on paper stuff where one party controls the client, server, application, and storage (trust me e2e security.)

Telegram, Discord, Facebook Messenger, stop using that shit.

This should be the top comment.

I recall blogs from over 20 years ago, with blatant comment spam, where the blog author would respond to the comment spam individually as if it was real readers. Most didn't fall for that, but a few clearly didn't understand it.

I'm not sure LLMs deviate from a long term trend of increasing volume of information production. It certainly does break the little bubble we had from the early 1990s until 2022/3 where you could figure out you were talking to a real human based on the sophistication of the conversation. That was nice, as was usenet before spammers.

There is a bigger question of identify here. I believe the mistake is to go the path of: photo ID, voice verification, video verification (all trivially by-passable now.) Take another step further with Altman's eyeball thing, another mistake since a human can always be commandeered by a third party. In the long term do we really care that the person we are talking to is real or an AI model? Most of the conversations generated in the future will be AI. They may not care.

I think what actually matters more is some sort of larger history of identify and ownership, matching to what one wishes to (I see no problem with multiple IDs, nicks, avatars.) What does this identify represent? In a way, proof of work.

Now, when someone makes a comment somewhere, if it is just a throw away spam account there is no value. Sure, the spammers can and will do all of the extra stuff to build a fake identity just to promote some bullshit produce, but that already happens with real humans.

Certainly the ICC would be using something more secure than Microsoft hosted email?

Wow, I had no idea he had died. This undersells what a critical influence he had in the direction the early internet took by setting a bar on what you could just post online (as an American.) Wikileaks got a lot of publicity, but Cryptome probably was much more important. If tech companies had more John Youngs and less sales people eager to sell censorship machines to dictators we would be living in an utterly different world today...

It's actually not that different than talking with employees, however, the LLMs still have very significant shortfalls (which you know about after using them a lot.)

If a manager doesn't know anything about what their employees are working on, they are basically fucked. That much holds up with LLMs. The simple stuff mostly works, but the complex stuff isn't going to pan out for you, and it will take a while to figure out that's the direction you went in.

I find now I quickly bucket people in to "have not/have barely used the latest AI models" or "trolls" when they express a belief current LLMs aren't intelligent.

If you are in a developing country and making $500-$1000 a month doing entry level coding work then $200 is crazy. On the other hand, your employment at this point is entirely dependent on your employer having no idea what is going on, or being really nice to you. I've also heard complaints from people, in the United States, about not wanting to pay $20 a month for ChatGPT. If the work you are doing is that low value, you probably shouldn't be on a computer at all.

"A well-designed self-service GUI" doesn't exist once the company is optimizing for the opposite of what your are trying to do (cancel subscription, return item, file insurance claim, etc.)

The network effect also works in reverse. The collapse of a network can happen very quickly. Facebook wasn't just functionally better than Myspace. Myspace experienced two really bad things at once, uncontrolled spam and adding display ads, at the same time, in 2008. It was easy to get users to just lose their daily Myspace habit and replace it with Facebook.

LinkedIn is unlikely to be mismanaged as poorly as Myspace was but there will be openings for competitors. Myspace and Facebook were unique because they were both very interchangable at that point, although Myspace was kind of like the "public" internet while Facebook had already dropped the edu email requirement, it was still heavily skewed college educated at that point.

Also worth pointing out, the edu email addresses requirement for Facebook likely did a lot for keeping their early network clean of bots and spam at a minimum cost. LinkedIn, on the other hand, basically hijacked their user's address books and sent out email impersonating those users, meanwhile ignoring unsubscribe requests and spam complaints. Which certainly sounds like something someone would end up in jail for doing (they did get in trouble for this.)

It's also not a reliable source of funding as some European open source projects have learned.

This is funny, I've been running remote companies for mostly the last 20 years (with physical offices during some of the time.) When I first saw the coworking concept and the inside of a WeWork I thought none of these people are working, they are just screwing around.

Thinking about what companies have the adtech infrastructure to buy Chrome and make enough money the justify a bid high enough than Google just shutting it down: Microsoft, Meta/Facebook, Bytedance?

I don't see a real need for Chrome. The stuff they've done to break Adblocking makes it pretty much a dead project today. Web browser development should be open source and not for profit. There is a fair argument that it has been because of Google's funding. There's a strong argument that Chrome has existed to further Google's business and at a minimum protect it's business and ensure third parties didn't hijack all of their PPC revenue in the early days.

It is easy to foresee an outcome here where someone politically connected gets a hold of Chrome and does a lot of crap they shouldn't. The worst case outcome is unrelated to any of this, and something where we end up with government mandated garbage in a web browser. It is very possible that DRM and biometric age verification, and who the fuck knows what else thanks to AI, could be required either by the US or EU, and kill the open source web browser. That's worse than anything Google did.

There are companies who have actually done this, like the whole Avast Jumpshot debacle. They aren't the only ones.

Another problem is you have no transparency with what ingredients restaurants are really using. Even if you know today, with certainty, you can't be sure tomorrow. Usually small chains will start out very high on the quality scale for suppliers and ingredients, and then swap them out as they scale/suddenly need to not lose money.

Food is tricky because once you have food processed or prepared, and you don't have the raw ingredients, you really have no fucking clue what you are ingesting. Most people act, and assume that they do.