EncoreBubble

@karpathy In terms of the atrophy, its also more like a mental reluctance to connect at the syntax level. A feeling of ugh, you just do this, I don't feel like it.

@karpathy The fun aspect. Yes. I definitely feel I can enjoy the fun parts and not get bogged down in the drudgery. Long term I do worry about the atrophy in case there is an incoming enshittification where one player wins and then the best models became prohibitively expensive.

A few random notes from claude coding quite a bit last few weeks. Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December. i.e. I really am mostly programming in English now, a bit sheepishly telling the LLM what code to write... in words. It hurts the ego a bit but the power to operate over software in large "code actions" is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do. This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks. I'd expect something similar to be happening to well into double digit percent of engineers out there, while the awareness of it in the general population feels well into low single digit percent. IDEs/agent swarms/fallability. Both the "no need for IDE anymore" hype and the "agent swarm" hype is imo too much for right now. The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk, in a nice large IDE on the side. The mistakes have changed a lot - they are not simple syntax errors anymore, they are subtle conceptual errors that a slightly sloppy, hasty junior dev might do. The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. They also don't manage their confusion, they don't seek clarifications, they don't surface inconsistencies, they don't present tradeoffs, they don't push back when they should, and they are still a little too sycophantic. Things get better in plan mode, but there is some need for a lightweight inline plan mode. They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines. They still sometimes change/remove comments and code they don't like or don't sufficiently understand as side effects, even if it is orthogonal to the task at hand. All of this happens despite a few simple attempts to fix it via instructions in CLAUDE . md. Despite all these issues, it is still a net huge improvement and it's very difficult to imagine going back to manual coding. TLDR everyone has their developing flow, my current is a small few CC sessions on the left in ghostty windows/tabs and an IDE on the right for viewing the code + manual edits. Tenacity. It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later. You realize that stamina is a core bottleneck to work and that with LLMs in hand it has been dramatically increased. Speedups. It's not clear how to measure the "speedup" of LLM assistance. Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion. Leverage. LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage. Fun. I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part. I also feel less blocked/stuck (which is not fun) and I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building. Atrophy. I've already noticed that I am slowly starting to atrophy my ability to write code manually. Generation (writing code) and discrimination (reading code) are different capabilities in the brain. Largely due to all the little mostly syntactic details involved in programming, you can review code just fine even if you struggle to write it. Slopacolypse. I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements. Questions. A few of the questions on my mind: - What happens to the "10X engineer" - the ratio of productivity between the mean and the max engineer? It's quite possible that this grows *a lot*. - Armed with LLMs, do generalists increasingly outperform specialists? LLMs are a lot better at fill in the blanks (the micro) than grand strategy (the macro). - What does LLM coding feel like in the future? Is it like playing StarCraft? Playing Factorio? Playing music? - How much of society is bottlenecked by digital knowledge work? TLDR Where does this leave us? LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.

@flaviocopes Lol yeah I really want to try it but also don't want to install it on main computer. But then if it's not on main computer, how is it going to help me in my real life? Need to figure out how to thread that needle. Or wait till someone else does.

I'm not sure who needs to read this but whatever you do ***do not install Clawdbot on your main computer*** especially if you don't know what your doing (something I assume only people in this category would do)

@kepano @0xdapper_ @thekitze Yes! Settings search or organization. Something is off about that, I find it harder than usual to find a setting in obsidian vs say vscode. And part of it is plugins vs core, but vscode has settings search. So maybe search is the missing part.

@0xdapper_ @thekitze I think adding settings search is one of the most important issues we need to address because half of people's problems are things Obsidian already has a setting for (c.f. thread) although I wonder if you would have searched for settings in this case? x.com/kepano/status/…

I love markdown but obsidian is not it

@MatSilva @greenwafflez @kepano @thekitze I just change the shortcuts. Cmd-p to search all files. Cmd-shift-p for command pallette, same as vscode.

@greenwafflez @kepano @thekitze so yes. i was holding it wrong. weeeeeeeeeeeeee

@Mooglet1 @charise_lee Yeah it makes them look off beat but that's not what they were dancing to.

@charise_lee Also that’s not the type of music they dance to. It’s old 60s music.

What in the (WHITE) world are we looking at here? Also what’s the WHITE Powder on the floor?

@leerob However I have been experimenting with using ai to document functionality in a new large codebase into obsidian as needed. I keep this separate from my notes because it's not "notes", it's "the textbook" 😂

@leerob Yeah I use obsidian and have tried using ai to write notes but I think the best notes for me are the absolute minimal diff against what's already in my head. What was "news to me". And ai can't know that. It writes too much and ... I ain't reading all that.

I'm happy with how I read email, write todos, keep notes. Maybe it's just me, but trying to throw agents and other products in here hasn't stuck. Feels like if anything I have went back to more basic things (just Apple Notes) versus some fancy optimizations. What am I missing?

@leerob Another similar one I'm experimenting with is saving Ralph like context to a file, so I can pick it up in a new agent.

@leerob Nothing too clever. One eg is /whereami to pick up where I left off from previous time I worked on this branch. Reads the branch name, uses jira mcp to get ticket, then checks the diff against main to tell me which acceptance criteria I have completed or not.

Rules, commands, MCP servers, subagents, modes, hooks, skills... There's a lot of stuff! And tbh it's a little confusing. Here's what you need to know (and how we got here).

@BrandonLuuMD What does gargling do, in general? Why gargle instead of say drinking a cup of green tea?

The window after exposure is short. Act quickly and you may avoid getting sick altogether. Even if you do get sick, these protocols can reduce severity and shorten illness duration. For the full evidence-based protocols, here’s the complete article: brandonluumd.substack.com/p/doctors-guid…

Someone just coughed on you. You were on a crowded flight. Now your coworker is sick. The window to prevent infection is narrow. Here’s exactly what to do in the critical hours after a high-risk exposure 🧵1/13

@GaryWade_Writes @parody_ceo It's literally in his name... Do they think Parody is an indian name?

@parody_ceo The subtly of satire is lost to some people.

Hired an employee He started off great for 18 months Then asked to miss work for a day Said his "sister was getting married" and he needed to cut out at noon on Friday Uh, sure dude. We've all tried that excuse before. I denied his PTO request. A few months later, he found a new job and left. I was relieved. Dodge a major bullet there. No doubt he was about to request another day off for a "grandparent's funeral" within a year or two.

@bethanyjbabcock Regular bread is just flour salt yeast and water. In the US, sugar and a bunch of preservatives. Here's standard wheat bread, some skippable non vegan things. But that said, let me tell you about the number of times my husband has bought vegan cheese!

@encorebubble No it’s made with milk butter and eggs

I don’t see the point of having the option to choose a female uber driver. What we really need is an option for female grocery shopper. These substitutions are head scratchers.

@bethanyjbabcock Isn't .. most bread vegan?

“They were all out of whole wheat sandwich bread in the store brand you wanted so I got you some other random product by the same brand instead of just a different brand of bread”—- what am I doing with vegan bread?

@DeerwoodRealty Also people focus on the monthly payment, so they will pay more than they should inflating prices even worse. Basically they will be renting with more steps.

So just working this out on the fly. The default 50 year mortgage talking point seems to be, “it’s just an option, you can still pay it back early”. Yes. The issue that I think most people have is that houses should not be so unaffordable that the 50 year option is needed. Maybe a better way to put it “Housing is too expensive” “Best I can do is 50 yr mortgage” The “fix” doesn’t align with the issue

@SaffronHawk @KobeissiLetter But NO FOOD! Food is bad, we have enough fatties.

@KobeissiLetter At this rate, by Monday he’ll announce free houses and debt forgiveness.

So far this weekend: 1. Trump announces 50Y mortgages 2. Trump announces $2,000 stimulus checks 3. Trump says insurance money should go to the people 4. Trump says tariffs will begin paying down US debt 5. Trump says no more money to insurance companies 2026 is going to be a wild year.

@JulieChangRE I wouldn't marry this person in the first place. But assuming I married them, that would imply at some point they were not like this, and something changed. So then I'd make them go to therapy and if it doesn't happen, I'd walk away.

Company: get a signature on delivery as this is a brand new MacBook pro UPS: Let me yeet this box over the gate of this apartment building. Oh it landed on the trash can, that counts as signature.

@dgsommersmkts Right from the 1960s. But your post said a clear difference from 2000 and I'm unclear what you meant by that because the graph shows acceleration of debt in both cases.

@encorebubble This is not normal, no. You can see an exponential rise in debt from the 1960s on. It's pretty clear.

Chair Powell: here is another clear difference between now and Year 2000. Sincerely, A Very Concerned Citizen