Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Future of AI Software Development

martinfowler.com

Post reply on HN
Print view View original

81–90 of 156 posts

Re: The Future of AI Software Development

#81

Post by heathrow83829 »

simonw wrote: post #9

> LLMs are eating specialty skills. There will be less use of specialist front-end and back-end developers as the LLM-driving skills become more important than the details of platform usage. Will this lead to a greater recognition of the role of Expert Generalists? Or will the ability of LLMs to write lots of code mean they code around the silos rather than eliminating them? This is one of the most interesting questi…

isn't it just one more step up the hierarchy. 10 years ago most developers have forgotten how to code in machine language because you didn't need to know it. Now, we're jsut going one step higher.

Re: The Future of AI Software Development

#83

Post by nyrikki »

lambda wrote: post #39
chadash wrote: post #11

> Will LLMs be cheaper than humans once the subsidies for tokens go away? At this point we have little visibility to what the true cost of tokens is now, let alone what it will be in a few years time. It could be so cheap that we don’t care how many tokens we send to LLMs, or it could be high enough that we have to be very careful. We do have some idea. Kimi K2 is a relatively high performing open source model. Peopl…

You don't even need to go this expensive. An AMD Ryzen Strix Halo (AI Max+ 395) machine with 128 GiB of unified RAM will set you back about $2500 these days. I can get about 20 tokens/s on Qwen3 Coder Next at an 8 bit quant, or 17 tokens per second on Minimax M2.5 at a 3 bit quant. Now, these models are a bit weaker, but they're in the realm of Claude Sonnet to Claude Opus 4. 6-12 months behind SOTA on something that…

It is crazy to me that it is that slow, 4 bit quants don't lose much with Qwen3 coder next and unsloth/Qwen3-Coder-Next-UD-Q4_K_XL gets 32 tps with a 3090 (24gb) as a VM with 256k context size with llama.cpp

Same with unsloth/gpt-oss-120b-GGUF:F16 gets 25 tps and gpt-oss20b gets 195 tps!!!

The advantage is that you can use the APU for booting, and pass through the GPU to a VM, and have nice safer VMs for agents at the same time while using DDR4 IMHO.

Re: The Future of AI Software Development

#84

Post by nyrikki »

cowmix wrote: post #54
lambda wrote: post #39

Earlier quoted context omitted.

You don't even need to go this expensive. An AMD Ryzen Strix Halo (AI Max+ 395) machine with 128 GiB of unified RAM will set you back about $2500 these days. I can get about 20 tokens/s on Qwen3 Coder Next at an 8 bit quant, or 17 tokens per second on Minimax M2.5 at a 3 bit quant. Now, these models are a bit weaker, but they're in the realm of Claude Sonnet to Claude Opus 4. 6-12 months behind SOTA on something that…

If you don't mind saying, what distro and/or Docker container are you using to bet Qwen3 Coder Next going?

I can't answer for the OP but it works fine under llama.cpp's container.

Re: The Future of AI Software Development

#85

Post by blibble »

chadash wrote: post #11

> Will LLMs be cheaper than humans once the subsidies for tokens go away? At this point we have little visibility to what the true cost of tokens is now, let alone what it will be in a few years time. It could be so cheap that we don’t care how many tokens we send to LLMs, or it could be high enough that we have to be very careful. We do have some idea. Kimi K2 is a relatively high performing open source model. Peopl…

> And it's hard to imagine that the hardware costs don't come down quite a bit.

have you paid any attention to the hardware situation over the last year?

this week they've bought up the 2026 supply of disks

Re: The Future of AI Software Development

#86

Post by markoa »

One thing that I'm sure of is that the agentic future is test-driven. Tests are basically executable specs the agent can follow and verify against.

When we have solid tests, the agent output is useful and we can trust it. When tests are thin or missing, the agents still ship a lot of code, but we spend way more time debugging and fixing subtle bugs.

Re: The Future of AI Software Development

#87

Post by varispeed »

crystal_revenge wrote: post #41
embedding-shape wrote: post #19

Earlier quoted context omitted.

> a near-frontier model Is Kimi K2 near-frontier though? At least when run in an agent harness, and for general coding questions, it seems pretty far from it. I know what the benchmarks say, they always say it's great and close to frontier models, but is this other's impression in practice? Maybe my prompting style works best with GPT-type models, but I'm just not seeing that for the type of engineering work I do, wh…

I’ve been running K2.5 (through the API) as my daily driver for coding through Kimi Code CLI and it’s been pretty much flawless. It’s also notably cheaper and I like the option that if my vibe coded side projects became more than side projects I could run everything in house. I’ve been pretty active in the open model space and 2 years ago you would have had to pay 20k to run models that were nowhere near as powerful.…

Depends what you see as flawless. From my perspective even GPT 5.2 produces mostly garbage grade code (yes it often works, but it is not suitable for anywhere near production) and takes several iterations to get it to remotely workable state.

Re: The Future of AI Software Development

#88

Post by lambda »

nyrikki wrote: post #83
lambda wrote: post #39

Earlier quoted context omitted.

You don't even need to go this expensive. An AMD Ryzen Strix Halo (AI Max+ 395) machine with 128 GiB of unified RAM will set you back about $2500 these days. I can get about 20 tokens/s on Qwen3 Coder Next at an 8 bit quant, or 17 tokens per second on Minimax M2.5 at a 3 bit quant. Now, these models are a bit weaker, but they're in the realm of Claude Sonnet to Claude Opus 4. 6-12 months behind SOTA on something that…

It is crazy to me that it is that slow, 4 bit quants don't lose much with Qwen3 coder next and unsloth/Qwen3-Coder-Next-UD-Q4_K_XL gets 32 tps with a 3090 (24gb) as a VM with 256k context size with llama.cpp Same with unsloth/gpt-oss-120b-GGUF:F16 gets 25 tps and gpt-oss20b gets 195 tps!!! The advantage is that you can use the APU for booting, and pass through the GPU to a VM, and have nice safer VMs for agents at th…

Yeah, this is an AMD laptop integrated GPU, not a discrete NVIDIA GPU on a desktop. Also, I haven't really done much to try tweaking performance, this is just the first setup I've gotten that works.

Re: The Future of AI Software Development

#89

Post by noelsusman »

crystal_revenge wrote: post #53
neebz wrote: post #23

Earlier quoted context omitted.

I've faced the same but my conclusion is the opposite. In the past 6 months, all my code has been written by claude code and gemini cli. I have written code backend, frontend, infrastructure and iOS. Considering my career trajectory all of this was impossible a couple of years ago. But the technical debt has been enormous. And I'll be honest, my understanding of these technologies hasn't been 'expert' level. I'm 100%…

A lot of people aren’t realizing that it’s not about replacing software engineers, it’s about replacing software . We’ve been trying to build well engineered, robust, scalable systems because software had to be written to serve other users. But LLMs change that. I have a bunch of vibe coded command lines tools that exactly solve my problems, but very likely would make terrible software. The thing is, this program onl…

This is the thing a lot of skeptics aren't grappling with. Software engineering as a profession is mostly about building software that can operate at scale. If you remove scale from the equation then you can remove a massive chunk of the complexity required to build useful software.

There are a ton of recipe management apps out there, and all of them are more complex than I really need. They have to be, because other people looking for recipe management software have different needs and priorities. So I just vibe coded my own recipe management app in an afternoon that does exactly what I want and nothing more. I'm sure it would crash and burn if I ever tried to launch it at scale, but I don't have to care about that.

If I was in the SaaS business I would be extremely worried about the democratization of bespoke software.

Re: The Future of AI Software Development

#90

Post by mattmanser »

markoa wrote: post #86

One thing that I'm sure of is that the agentic future is test-driven. Tests are basically executable specs the agent can follow and verify against. When we have solid tests, the agent output is useful and we can trust it. When tests are thin or missing, the agents still ship a lot of code, but we spend way more time debugging and fixing subtle bugs.

Great, so now I have to design the API for the AI, think of all the edge cases without actually going through the logic, and then I'm invariably going to end up with tests tightly coupled to the implementation.
Post reply on HN