Which I use for years, and just because of lack of a better alternative (minimal, OSS) I actually quite "hate" it.:
It has no rules, like "Allow a file with THIS name on THAT path", the biggest complain of every simplewall user is its biggest flaw and annoyance: It asks you a million times to allow or deny the "same" app, after its checksum changed. So all the Windows apps that auto update always pop up again and again.
A coding agent autonomously compiling and testing an exe with network access is impossible. I have to turn off simplewall, because it would constantly block the executable after recompilation.
I'm a simplewall user and what you describing is not my biggest flaw or annoyance - I simply don't upgrade windows (why would calculator need updates??) and after clicking few denies, its get very quiet. Been this way for 2 years now on Win 11.
doesn't it also threatens most servers currently not on planet earth? like on chinas space station? i guess ISS's hardware is too old and UEFI didn't exist back then, so they are probably safe.
i filed my swiss taxes for 2025 in 2026 (april) by dumping everything (local tax law, tax guide, my and my wife's documents, bank statements, income statements, etc.) into a folder and asking claude to fill it out. i had nothing to fix. submitted it.
I did mine and my wife's. Mine is really easy but my wife's situation is a bit harder, because she both works as an employee and as a freelancer. Claude was able to make sense of the excel spreadsheet sent by the accountant, and to translate it into what was expected on the tax portal.
It beats the tax advisor we hired two years ago, who made a significant mistake.
This falls into the obscure ("not important") but not difficult part of problems, that has an extreme long tail, compared to exceptionally difficult and important.
indeed as the author mentioned, LLMs can greatly help in areas where there is a long tail of not so important, easy to solve problems, that humans just don't have the time or priority to focus on. But combine this long tail of problems that can be solved: accumulated this might still be very beneficial as a sum of things.
who tf uses prompting to "pretty please don't cheat on this"? the best practices for ages (in terms of ai) is to separate the eval from the test code/agent.
another best practices every single solution using LLMs/agents should implement is "never trust the llm".
> who tf uses prompting to "pretty please don't cheat on this"?
People who don't understand how LLM's work. Kids, teachers, your next door neighbor. CEOs and government officials.
I am not trying to argue that the author doesn't understand how LLM's work; they clearly do. Their prompt was written in a way that leaves those assumptions implicit, in a similar manner to those non-technical folks using LLMs every day.
> the best practices for ages...
We cannot expect the entirety of humanity to understand / use LLM best practices. We need to figure out how to ensure LLMs operate as the majority of humanity expects despite how they are implemented, and until we can do that, do our best to prevent them from being used in an unsafe manner by the nontechnical population of humanity.
for claude, it's response headers contain available usage, you can get it there. My deepseek harness plugin auto stops asking things when I am at 80% to allow inflight prompts to finish.
Please challenge me and explain why this is a break through worth the headline they use for their own work.
How I understand it, without really reading into it:
- Thomson Reuters did not "create" a frontier model, they gave some money, maybe a bit of their data to Imperial College London, and took Alibaba's Qwen 3.6 35B A3B Model for a basic fine tune.
- "They" (some undergrads at Imperial College London) used an existing ablation framework to undo some of the topic alignment of the original model.
- "They" fine tuned on some domain knowledge trying to preserve general knowledge - here maybe, just maybe some data came from Thomson Reuters.
As a result: one of 100's of Qwen 3.6 35B A3B sparse model fine tunes, just for publicity, to write overstating headlines like "X created their own frontier model"
That's pretty wild that a Ctrl-F of "qwen" or "china" or "alibaba" on the thomson reuters announcement URL turns up no results. They want to portray it like they developed this thing from scratch, when that is by no means what actually happened.
They didn’t say they created it no? I don’t think any of those points matter. It’s a business announcing a new product. Why does it matter if they built it internally or used external teams? I find strange to focus on those details
Hyperpersonalized code to its rescue. With the qwen3.8-27b release tomorrow and the advent of vibe coding and the improvements into proper agentic engineering we are moving towards hyperpersonalized software.
This means we don't need a centralized maintainer for smaller tools. the facebook ad blocking cat and mouse game is something I cna very well imagine to just be something my computer would solve in real time, updating the uBlock filter/userscript to adapt to any fb changes in near real-time.
Sadly they seem to not be releasing a sparse 35b A3b or anything inbetween "too large to host for mortals" and "fits into a consumer rtx". Probably not to eat away their profits on their API serving. 120b - 300b is a dead space right now, very few good releases in that size range. (I know there are, but the big labs aren't releasing stuff here)
Hyperpersonalized code to its rescue. With the qwen3.8-27b release tomorrow and the advent of vibe coding and the improvements into proper agentic engineering we are moving towards hyperpersonalized software.
This means we don't need a centralized maintainer for smaller tools. the facebook ad blocking cat and mouse game is something I cna very well imagine to just be something my computer would solve in real time, updating the uBlock filter/userscript to adapt to any fb changes in near real-time.
It has no rules, like "Allow a file with THIS name on THAT path", the biggest complain of every simplewall user is its biggest flaw and annoyance: It asks you a million times to allow or deny the "same" app, after its checksum changed. So all the Windows apps that auto update always pop up again and again.
A coding agent autonomously compiling and testing an exe with network access is impossible. I have to turn off simplewall, because it would constantly block the executable after recompilation.
reply