The best software is both the right software, and high quality. It’s bloody hard to do!
So, we do requirements gathering, and try to get the details right (TDD, continuous delivery etc).. it’s surprising to see Gas Town do none of this and optimistically hope agents will converge on good software by just throwing tokens at a wall and hoping it sticks.
So anyway, that’s what they’re ignoring. What are they actually doing???
It’s all for themselves.
Gas Town’s “product” is the warm fuzzies it gives people with money to burn, warm fuzzy feelings of being “at the frontier”. It’s a luxury product for nerds, and the only ones making money or selling anything are the big labs. There’s zero output or benefit to society because that’s simply not the point.
I think Gastown is truly special, but I wanted something more focused on learning as I think that's the real bottleneck. So I built AgenC to make it trivial to roll learnings back into your Claude.
I was a NNW user for years and it's why I eventually built my own news reader. NNW had a lot of great features and I wanted to mostly keep them. You might find that NewsBlur takes a similar path but with a different set of opinions.
Personally I keep it syncing off TTRSS for filtering and automatic actioning on certain feed entries, but that aint everyone's cup of tea. I'd like to think NNW at least covers most people's use cases whether standalone or relying off another service to aggregate.
RSS’ death is real - 15 years ago, almost every news site had a RSS feed, some had several ones. Today? RSS feed is rare.
So if you want to make news feed from news sites, you have to use parsing their html code, and ofc everybody has its own structure. JS powered sites are painful ones.
15 years ago, almost every news site had a RSS feed, some had several ones. Today? RSS feed is rare.
It may be a reflection of where you get your news.
New York Times, Washington Post, Wall Street Journal, Radio Free Europe, Mainichi, and lots of other legitimate primary source Big-J journalism news sites have RSS.
Rando McRepost's AI-Generated Rehash Blog? Not so much.
Seriously. I've been updating NewsBlur with all the pet features people have wanted for years and I'm finding that it's even more enjoyable now with all those AI features built in. Daily briefing, ask AI, story clustering, all of these are AI-flavored improvements to RSS and it's so relaxing to open up my river of news and scroll through all the good stuff without feeling a gross algorithm surfacing endless outrage.
I read plenty of X as well as scroll through various social media apps and nothing comes close to how great RSS feels to read.
As someone clumsy, I'm so grateful that my MacBook Air can take a beating. It has one slight dent of about 1mm in the 4 years I've had it and I definitely drop it or knock it off a desk or something a few times a year.
I'll take the extra weight of aluminum (0.3lb, 130g). Yes, someone might say the ThinkPad X1 Carbon is 14", but the 13" MacBook Air actually has a 13.6" screen.
If I were in the market for a PC laptop, I'd definitely take a look at the ThinkPad X1 Carbon, but I'm also not worried about the weight of my MacBook Air. The X1 Carbon Intel ones are on sale right now since Panther Lake will be a huge upgrade coming soon, but even on clearance they aren't cheap. An X1 Carbon with 32GB RAM and 1TB storage (Ultra 7 268V, the cheapest one due to the sale) will cost $1,679 while a similar MacBook Air will cost $1,699 - and the M5 has 48% better single-core performance and 56% better multi-core performance (Geekbench). A 16GB/512GB (Ultra 5 225U) X1 Carbon is $1,538 compared to $1,099 for a MacBook Air - and the M5 has a 74% single and multi core advantage there.
Panther Lake might narrow the performance gap, but early indicators don't seem like that's the case. Even the top of the line Ultra X9 388H sees the M5 with a 36% single-core advantage while the Ultra X9 388H gets 3% faster multi-core. And I'm not sure the higher wattage "H" processors work for something like an X1 Carbon.
The highest non-H Panther Lake processor (Ultra 7 365) sees the M5 get 51% better single-core and 58% better multi-core. Maybe we'll see better, but it looks like Intel isn't closing the gap in 2026.
Does it? In my case, it was my father who dropped my mac but luckily everything was all safe with tis but a scratch. So perhaps that can be taken into factor as well that its more than one variable.
That being said, I am pretty clumsy but I have never dropped any hardware except a dumb phone which I threw out a lot and it was so small and tiny but it never had any problem.
And then one day I dropped it from top just a little bit and let it drop/slide inside my bag (like a cushion) and that day it died. I recently asked someone about it and turns out that its battery got inflated.
I suggest adding Stop & PreCompact hooks to your agent which give it the log and ask it to review its own actions in case it did anything unsafe or unexpected. It can check what it sees against what it remembers and tell you if anything stands out.
Or you could give the transcript and log to a new model and have that one do a review.. but either way the goal is to reduce your cognitive load.
Even cooler is when you notice you can have the model provide recommendations - and make its own plan to incorporate them :)
Have your agent do red/green TDD - its like double entry accounting, the tests and code mirror each other and the tests are an executable repository of docs of how the thing should behave, that helps immensely when you need to be an archaeologist and understand some deep corner of the system.
Code reviews are a cinch because if you get confused by the real code you can switch to reviewing the tests and vice versa.
Alignment is fragile; this is demonstrated by how complex society is. One minor perturbations like decreasing police oversight leads to wildly different outcomes on metrics we care about, like not having to be afraid of getting randomly murdered. Our demons must be put down constantly. Nothing about this will ever change, no matter how much we educate the children.
That said I think this is overblown. In 2026 a bunch of awful social media companies dominate the landscape and benefit from the bottom feeders populating feeds with generated content; that’s a big cause of the erosion of trust. In a healthier landscape there would be viable anlternatives and that’d allow experimentation - stuff like attestation of your posts’ provenance, and the like.
We’ll get there, I just think online is in the “robber baron” era, where a few monopolists have captured the market. But you can see the cracks emerging with open source and self hosting and EU regs and EU government procurement and talk of regulating algo feeds. It’s many years away, so your best bet for now is to avoid feeds, touch grass, hang in group chats, and don’t let yourself get too cynical.
I just migrated to Fastmail (on my domain), it’s fantastic. It works just like Gmail in every way I need, haven’t missed Gmail or Google Calendar one bit. It’s clearly made by people who know Gmail well and understand why it works the way it does. I thought it’d be a huge migration but it was actually boring. Search works, 20 years of emails just magically migrated over. Spam detection is better. Couldn’t be happier!
Accidentally typed gmail.com the other day, it took 4 seconds to load (Fastmail is instant) and when it finished loading there was an ad to try some paid Google service. Felt like a flashback to an abusive ex.
Some of the ideas here are codified in a simpler way in my TDD framework, CodeLeash (https://codeleash.dev).
The key difference is CodeLeash puts guardrails outside the model, as Claude Code hooks. It can’t forget or skip steps, it gets forced through a process. There are bypasses but they raise alarms that the model is also forced to review.
Putting guardrails outside the model forces deterministic process following. With a sufficiently capable process (TDD is a powerful way to build software) you can really scale up coding agent usage.
Cool but it is not a framework for working with AI, it is an _opinionated_ framework for building full-stack apps right? As in, I can't use any of it if I'm building, say, a Spark data processing pipeline. Or a ML framework. Or automation software that runs on custom processors.
The idea of "guardrails outside the model" is definitely appealing but I wonder if you can make it generalize well.
You’re right of course, the framework is pretty rigidly tied to full stack.
But the underlying idea I think has power - find a process you can codify enforcement of rather than telling models how they should do things.
Spec driven development probably means creating tooling to track how code maps to specs, for example.. and then working out how to manage that as data. Then you can query the data to confirm all mappings between code and specs. That gets you out of the business of nicely and repeatedly asking very expensive and undependable models such queries :)
The TDD guard is the most isolated part of the whole thing, you could just take the files involved and ignore the rest. Maybe 6-7 files all referenced from the .claude/settings.json hook config.
First is understanding what to build.
Second is getting the details right.
The best software is both the right software, and high quality. It’s bloody hard to do!
So, we do requirements gathering, and try to get the details right (TDD, continuous delivery etc).. it’s surprising to see Gas Town do none of this and optimistically hope agents will converge on good software by just throwing tokens at a wall and hoping it sticks.
So anyway, that’s what they’re ignoring. What are they actually doing???
It’s all for themselves.
Gas Town’s “product” is the warm fuzzies it gives people with money to burn, warm fuzzy feelings of being “at the frontier”. It’s a luxury product for nerds, and the only ones making money or selling anything are the big labs. There’s zero output or benefit to society because that’s simply not the point.
reply