KolibriFly

Pleasant reminder that building good architecture for agents just ends up mirroring time-tested software engineering practices.

Bro, it really just downloaded an executable off the internet because it could. The only humans it 'used' here are the cheat developers who open-sourced their code

Its funny watching Open AI and Anthropic play the role of IP defenders while quietly sliding Deepseek's KV cache research into their minor releases to salvage their unit economics

An ai wrote a post explaining how to give an ai hands , and half the comments spotted the prose before even reading the code

peak tech blogging right here

Disliking clunky autocomplete is very different from going out of your way to replace Word with Libre Office. The average office worker will tolerate ten bad assistants before they voluntarily deal with ODF formatting quirks in a corporate environment

Wait until they find out people use compiler output to write better compilers without asking for permission

Glad to hear they are prioritizing evaluation right from the start. Usually people just throw together a rag pipeline on the knee and then judge the metrics by eye, skimming three responses in the terminal

You should feel ashamed of an unreviewed 1000-line PR that crashed prod, not an autocomplete saving you a couple hours of your life

That's docker, man. Tomorrow they're gonna add limits on sandbox runs without a premium account too

Squashing commits pretty much turns the whole "green carpet" into meaningless noise

Non-techies avoid the CLI like the plague and don't want to mess with .gitattributes for text files. But building a whole web service with a paid subscription just for that is total overkill when free desktop GUIs like Sublime Merge or VS Code exist

It's pretty ironic to pitch protection against AI accusations with a post that was completely generated using Claude prompts

With that logic you could call SQL injections a natural feature of database management systems. If a general purpose system starts dropping tables or messing up numbers in a report just because that string was in the text it read, that system isnt worth a damn in the enterprise sector

A court should not quietly create a general duty to inspect everyone's private files because a provider could theoretically detect illegal content

Yeah, that is pretty funny considering Blackwell isn't just a chip but a whole ecosystem with a fifth-gen interconnect and huge cooling racks...

Another important thing to keep in mind here is the software. Nvidia has been optimizing their software stack for decades to squeeze every clock cycle out of the hardware under any limits. TensorRT alone does straight up magic. Apple is just starting out with MLX, and their hardware is often idling not because it is worse per watt but because the compiler does not know how to optimally load the ALU blocks for specific graphs yet

A lot of the article is about who controls the tool, who is forced to use it and who absorbs the cost when it produces bad output

I suspect "same price, three days instead of three months" is a very real win for plenty of businesses

I think the argument is that AI changes the cost of creating that groundwork, not the definition of it

The real trick is recognizing when "disposable" code has quietly become infrastructure

AI makes it cheaper to create working fragments. It does not automatically make those fragments part of a maintainable system. In practice it may make the canonization step more important not less

Everything depends heavily on the segment. Open-weight models already look good enough for a ton of internal tasks right now but there is always gonna be that small percentage of workloads where the last few percent of quality are totally worth the money.

I feel like the market is just gonna become way more mixed. Not like "everyone is switching to open" or "everyone is staying on frontier" but a mix of multiple models for different scenarios

The weakest part of the article is that the forecasts up to 2029 assume the current market structure will barely change. In three years literally everything can change, like prices, models, hardware, and how we actually use LLMs

That is exactly why big clouds never put all their eggs in one basket, even if it is a super cheap and cold basket. The cost of protecting and backing up network lines for an isolated island quickly eats up any benefits from geothermal energy. Physical security for terabit lines is way more expensive than air conditioning these days

Iceland and Norway are part of the EEA so the AI Act and GDPR will reach them just like Germany or France. Running away there from regulators makes no sense. But running from bureaucracy to get land permits and substation connections - yeah maybe municipalities work faster there

I mostly agree, though I'd phrase it slightly differently: software is becoming less of a moat by itself but not necessarily less important

The people saying "anyone can build software now" often seem to mean "anyone can generate code now"... which is not quite the same thing

Yeah, but maybe only for the visible (productized) layer of software

I feel like this is way too binary. I don't have to write every line of code myself to understand the system. I don't write my own compiler HTTP stack or database either

It's more about the level of abstraction. If AI handles 80% of the grunt work and I spend my time on architecture and reviews that's still a win

The funniest thing would be if in a couple years LLMs just end up being another checkbox next to PostgreSQL and Kubernetes