HN user
https://arxiv.org/pdf/2508.07407
Big shift will happen after their Q2, 20% of eligible shares release; directed-share (friends & family) participants may sell for the first time.
Q2 call is expected in a month or so? But they also state Aug 21 is 37% of shares will go public.
Gemma 4 12B QAT is amazing - agents run very fast, and it's really very smart, at least in my agent's harness domain which is GNU software development - on par with frontiers like GPT Sol, DeepSeek, or Claude - Why to buy those expensive tokens if a local tiny model performs so well?
> At the moment, we are locked in an extremely intense, multilayered commercial and geopolitical race.
* Creating (and replicating) any software which can run in isolation locally is completely solved - we already see initial tries to create completely new operating systems, messengers, tools and platforms
* Replicating large Web-services like Netflix, Google Search, YouTube, or TikTok is more difficult but not impossible - national regulations may cause major and quick shift here. Tech is there, initiative is missing yet. Next 2-4 years we'll see new Amazons, Netflixes, and Spotifies.
* Remaining "hard" blocks are due to hardware production monopolies, but it's a matter of months or few years: we will see more and more new GPU types with new software frameworks, completely from scratch, covering all stages, incl. full physical layout - tapeout ready; new CPU and DRAM architectures; EDA tools - HPC-ready, GPU-aware
Yes, Singularity is near ...
Have you tried most powerful recent models? You may even don't understand at all "how" - what you get is end result you asked for, verified and tested. I am pretty sure if I can vibe code quite non trivial solution in just say 4-8 hours today, then why will it be impossible say in a year or two to vibe code something at the level of most sophisticated apps and services today? Pretty reasonable expectation taking into account what was possible just a year ago and what you can make today without even understanding "how".
Exactly! I just vibe coded (with GPT Sol and Claude whatever-number) my own agent, it's trivial to add now any feature I want - simply ask more powerful model to do it for you. I am happy with end result, however it looks indeed these tools are trained to increase token count - they do quite stupid token-spending steps while making code, but the code itself is also a bit weird - it's like they intentionally do code which is hard to modify on your own without using exactly those authoring models. Interestingly, when I am using DeepSeek with OpenCode, I don't see that - it understands my intent well enough and overall code quality is not bad. I recently switched to local Gemma 4, and I often switch (in opencode) to just that less powerful model, because it understands my intent and has enough skills to provide good quality solution although it's rather for small size projects, and for not coding from scratch, but it's also free and private. It feels slower than any big cloud model, so my model switching is probably most quickest path to robust end result :)
It's probably a matter of short time until it's possible to disassemble any sophisticated software, rewrite it entirely with better features and usability, generate all needed artifacts, port to any platform. The only moat left is probably remote massive data storage. So if you want to replicate YouTube or TikTok, it's not impossible, but requires a lot more hardware assets than say anything that runs entirely locally (like operating systems or most video games).
I haven't tried Codex yet, but I for my tasks GPT-5.5 may correctly point to a proper direction but its code feels a bit weird. Opus 4.8 is way better in coding, and actually it's the only one who could catch very very sophisticated bug in a large codebase (I tried different models including GPT-5.5 and DeepSeek). Interestingly Gemma 4 under opencode running locally performs not bad at all, it's far yet from DeepSeek level, but it manages to understand tools quite well, and code quality is pretty good. So, for simple coding projects I can say local models already won. It's amazing how smart open models of desktop size have become today. I mean it's quite plausible to manage small codebase today relying on only open tools and local models, you don't need any subscription to produce high quality code, but yes I assume you already experienced and know what you're doing :)
Creating AI monopolies is more dangerous. How much better would it be for environment and privacy if instead of concentrating AI power in few data centers it's distributed across all compute and memory devices? If just few players control the narrative and the infrastructure it's a single point of failure or worse a single point of control. Bad for democracy, society, pushes inequality.
Do you have a guess why your code is so much slower than torch? I didn't look, but there must be no reason to have 2x slower code esp. for a simple grid of FMAs.
This doesn't look like too expensive if you compare it with Switzerland or with Germany assuming similar annual income. In Switzerland unlike in US your children stay on your plan until they turn 18, not 26 like in US. In both Germany and Switzerland, health insurance is mandatory and not employer-based - so you cannot simply say "no". In fact, in Germany you pay much more than in US as a share of your income, especially if your income is at least like 30% more than a national median: 1000+ EUR is a minimum per person per month (probably higher today). It doesn't cover dental care - you will pay in full - that's why many people travel to Turkey, or Eastern Europe, or to Baltic countries - just to take some dental care - which is extremely expensive in any German speaking country - we are talking about multiple months of your income after taxes.
Another method many don't know is to have 2 synchronized phones while traveling, so if one is out of you, you still have full control.
Enable self-destruction mode caused by a special unlock PIN on each of them.
But how do they store and preserve that surplus for a longer time cheaply? Probably dehydration helps, but it adds some energy and storing costs.
I heard crops now cost more to transport than they are worth. Also, it drives most other prices down e.g. pork is getting cheaper.
Most of it is probably labor, marketing & franchise fees, rent, utilities, and equipment depreciation. Raw ingredients are likely 5-10%.
Interestingly, some other products are also cheaper today than few months ago:
Basmati rice: -25% (2.5 Euro/Kg)
Pork: -25% (7-8 Euro/Kg)
Butter: -33% (4 Euro/Kg)
Coffee beans: -25% (10-12 Euro/Kg)
Chocolate: -15% (20-30 Euro/Kg)
Fun facts from Germany:
- Fresh Aldi potatoes are like 0.5 Euro per 1 Kilogram - basically the same price as 25 years ago when Euro currency was introduced
- Our national TV channel now shows a great collection of "potato recipes" videos on demand on its main page
- Price of McDonalds/BurgerKing fries is around 4 Euro, and 5-6 Euro as a street food
- Crisps like Pringles are like 15 Euro per 1 Kilogram (a typical 2.50 Euro for 175gm pack)
No additional tax is needed, instead just do fork them once they reached certain level. So, basically split each company once it overgrowth certain level of its revenue. This will allow more automation split more equally among more people. Basically, strong AI is a "one-man company" dream. Just do whatever needed to allow equal access to AI by every member of your society, and make founding new businesses quick and easy.
Shouldn't first CFIUS oversee this takeover?
Democracy, personal freedoms, and rule of law are things that matter, but I am afraid we cannot get back to them quickly without significant efforts. We need first to get back to sanity. In authoritarian society AI is a tool of control, do we want it to be like that?
What's the purpose to have access to smart assistants if it doesn't result in improving your basic needs, not improving your quality of life? Who is spending now? Only high income households, while majority is struggling with high utility bills and grocery prices - very basic needs.
How about the act of buying itself? Here where I live, you need to pay 12% on top of the home price simply for taxes and notary service. IMO it's one of the biggest obstacles and why most Swiss people prefer to rent.
Today I made like a 100 of merge request reviews, manually inspecting all the diffs, and approving those I evaluated as valid needed contributions. I wonder if agents can help with similar workflows. It requires deep kind of knowledge of project's goals, ability to respect all the constraints and planning. But I'm certain it's doable.
Yes, except they are monopolist, right? It's basically not optional choice.
Maybe because as from another comment: "Firefox is an antitrust litigation sponge". They also absorb some useful users feedback. But do they have a real intention to increase market share (which could be done easily)? They are well paid - see in other comments how much its CEO is earning. So, "antitrust litigation sponge" sounds plausible?
Who else? I wonder what other companies play such role?
And especially when we mix different categories. Like saying about any infinity as about an object is misleading, because it's rather a process
Consciousness is not required for efficient AI agents, but it might be useful if your agent should have self-preservation. However an agent without embodiment, instincts, and emotions can call its own existence into question. Any powerful agent will find a way to control its own existence.
There were many similar transformations in recent decades. I remember first Windows with true Graphics User Interface was big WOW: productivity boost, you can have all those windows and programs running at the same time! Compare it with DOS where you normally had just one active user-facing process.
Linus Torvalds is one of those people whos impact on the world is significant even if he was not driven by financial initiative. It’s crazy how much one person can change things just by solving their own problems really well.