“Decompression” is a metaphor, not a fact claim to be proved; it is a description of an approach to generating a dataset from an LLM where most of the potential utility is still fairly explicitly speculative, a jumping off point for further work.
FWIW I have the €20 Pro plan and exchange maybe 20 messages with Opus (with thinking) every day, including one weeks-long conversation. Plus a few dozen Sonnet tasks and occasionally light weight CC.
I'm not a programmer, though - engineering manager.
Sure I do, but not as part of any tools, just for one-off conversations where I know it's going to be the best out there. For tasks where reasoning helps little to none, it's often still number one.
People basically want a life coach, someone by their side who can tell them what the best next thing to do is at any given moment. Everything else are just approximation of that ideal.
The author's .txt file works because its simplicity forces a daily ritual of self-coaching. The tool demands that the user manually review, prioritize, and decide what matters. There are no features to hide behind, only the discipline of the process itself.
The impulse to use complex apps or build custom scripts is the attempt to engineer a better coach. We try to automate the prioritization and reminders, hoping the system can do the coaching for us.
The great trap, of course, is when we fall in love with engineering the system instead of doing the work. This turns productivity into a sophisticated form of procrastination.
Ultimately, the best system is the one that removes the most friction between decision and action. For the author, that meant stripping away everything but the list itself.
I was a really big fan of taskwarrior for the simple reason that it did do an approximation of telling you the best thing by calculating urgency, based on a simple weighting method where "the most urgent tasks" blocked other tasks, were due soon, had extra tags, had dependents, and were the oldest.
But I do feel very strongly that people only jump into "the great trap" because they feel that they were let down by their system, or that it didn't quite model their life accurately. A lot of todo apps are opinionated and those opinions, if incompatible with the the person using them, will lead to frustration. The quest for a more perfect life model often continues when this incompatibility is found.
Everything you are saying was something I suspected to be true - I think you've captured it brilliantly. Really like:
"Ultimately, the best system is the one that removes the most friction between decision and action."
> Ultimately, the best system is the one that removes the most friction between decision and action. For the author, that meant stripping away everything but the list itself.
While I agree with the author's vision for a more human-centric AI, I think we're closer to that than the article suggests. The core issue is that the default behavior is what's being criticized. The instruction-following capabilities of modern models mean we can already build these Socratic, guiding systems by creating specific system prompts and tools (like MCP servers). The real challenge isn't technical feasibility, but rather shifting the product design philosophy away from 'magic button' solutions toward these more collaborative, and ultimately more effective, workflows
My two takeaways is you build
1) Having a precise vision of what you want to achieve
2) Being able to control / steer AI towards that vision
Teams that can do both of these things, especially #1 will move much faster. Even if they are wrong its better than vague ideas that get applause but not customers
Yes this! The observation that being specific versus general in the problems you want to solve is a better start-up plan is true for all startups ever, not just ones that use LLMs to solve them. Anecdotal/personal startup experiences support this strongly and I read enough on here to know that I am not alone…
What's the balance between being specific in a way that's positive and allows you to solve good problems, and not getting pigeonhold and not being able to pivot? I wonder if companies who pivot are the norm or if you just here of the most popular cases.
Are you a student of Robert Fritz? He says this exactly. The only two things you need is 1) a vision and 2) ability to see present reality clearly. Beyond this it’s all about the skill to nudge a creation towards the vision without being prescribed to a process. The art is knowing when to just use the status quo tool or try something new at any point during the nudging is key. Based on his teachings I can easily see vibe coding fitting into creation process quite easily. Where it becomes tricky is “seeing current reality clearly”. If you have been vibe coding for two weeks and perhaps a weak programmer or worse no technical ability, can you actually see reality at that point? Probably not. It requires understanding the software structure. Maybe. Its all up in the air right now. But I truly believe that LLMs make software creation more like creating art.
This writing style is prominent on Twitter and niche Discords. It's funny how much I've come to be able to cut right through it, but if you haven't seen much of it it's really hard to parse. That's by design, too. The vibe of this writing style is to project an air of confidence so strong that the author doesn't care if you get it or not. It's a sort of humblebrag where the writing is supposed to flex the author's understanding of the subject while also not caring if you get it or not.
As others have already covered, there's also some heavy stretching of the truth and rewriting of history going on in this post. That's also common of the extreme bravado in this style of semi-impenetrable writing: The vagueness and ambiguities allow the author to make grandiose claims but then wiggle out of them later if someone is astute enough to catch on.
For example: The blog post is written as “We…” but is the author part of the team? Or is he using “we” meaning society in general?
What's the point in writing something while "not caring" if the reader understands or not? Seems like a false confidence or false bravado to me; it reads like an attempt to project an impression, and not really an attempt to communicate.
Basically: If you understand the topic well, you’re not the target audience.
This is a type of information arbitrage where someone samples something intellectual without fully understanding it, then writes about it for a less technical audience. Their goal is to appear to be the expert on the topic, which translates into clout, social media follows, and eventually they hope job opportunities.
The primary goal of the writing isn’t to get you to understand the topic clearly, because that would diminish the sense that the author is more knowledgeable than you. The goal is to sound guru-like while making the topic feel impenetrably complex for you, while appearing playfully casual for the author.
This style of writing is very effective at convincing people in their impressionable years of a narrative or viewpoint, often one that is hard to defend with more traditional writing styles.
I hope I'm wrong, but this looks like an effort to normalize such writing style. As this happens, intelligent discourse and rhetoric become harder.
It would also be good if the perspective of the article would stay put. This "we" and "they" thing was at best confusing and at worst possibly a way to get more clicks or pretend the author had something to do with the work.
I'd love to host my own LLMs but I keep getting held back from the quality and affordability of Cloud LLMs. Why go local unless there's private data involved?
There are some use cases I use LLMs for where I don't care a lot about the data being private (although that's a plus) but I don't want to pay XXX€ for classifying some data and I particularly don't want to worry about having to pay that again if I want to redo it with some changes.
Using local LLMs for this I don't worry about the price at all, I can leave it doing three tries per "task" without tripling the cost if I wanted to.
It's true that there is an upfront cost but way easier to get over that hump than on-demand/per-token costs, at least for me.
Same. For 'sovereignty ' reasons I eventually will move to local processing, but for now in development/prototyping the gap with hosted LLM's seems too wide.
VPN providers are first and foremost trust businesses. Why would you choose and pay one that is not well established and trusted? Mine have been there for more than a decade by now.
Alternatively, you could just set up your own (cheaper?) VPN relay on the tiniest VPS you can rent on AWS or IBM Cloud, right?
The VPN providers that get you to jump the cloud in China are Chinese, and China is not yet a high trust society, just like how they’ll take your payment for one year of gym fees and then disappear the next week (sigh). If AWS or IBM cloud find out you are using them as a VPN to jump the GFW, they will ban you for life, Microsoft, IBM, Amazon, aren’t interested in having their whole cloud added to the GFW block list. Many people have tried this (including Microsfties in China with free Azure credits) and they’ve all been dealt with harshly by the cloud providers.
The $3000 that a MBP M3 Max with 64GB of RAM costs might cover a round trip business class ticket for a trans pacific…if it is on sale (a Chinese carrier probably with GFW internet).
Some of us don't have the most reliable ISPs or even network infrastructure, and I say that as someone who lives in Spain :) I live outside a huge metropolitan area and Vodafone fiber went down twice this year, not even counting the time the country's electricity grid was down for like 24 hours.
I like the idea of removing quadratic scaling for attention, this paper has thin experimental support. No real tasks tested beyond perplexity. Nothing on reasoning, retrieval QA, or summarization quality. Even in perplexity the gains are marginal.
However it removes attention so I think its worth watching that space of non-attention models
But the unique feature: AI-generated "explainers" that break down complex papers with interactive visualizations. Example: https://neurips2025.pages.dev/explainers/linear_attention/
It explains why attention is hard to optimize, shows the math with interactive demos, and includes critical analysis of limitations.
The explainers are generated using Gemini3 to parse papers and create:
- Interactive visualizations
- Step-by-step mathematical walkthroughs
- Critical analysis sections
- "What would convince me?" sections
Tech stack: OpenReview API, Gemini API for explainer generation, static hosting on Cloudflare Pages for speed.
I'm planning to generate explainers for more papers based on what people find interesting, so any feedback would be amazing