Alex Pilon

1.4K posts

Alex Pilon banner
Alex Pilon

Alex Pilon

@xaptronic

some kind of growth hacker @shopify; prev. @tripadvisor / @travelpod - moonlight musician / audio hacker - data science wannabe - cats of course

Canada Katılım Temmuz 2008
515 Takip Edilen197 Takipçiler
Alex Pilon
Alex Pilon@xaptronic·
@nejatian What is the difference between buying "a Tesla" and buying TSLA?
English
0
0
0
87
Kaz Nejatian
Kaz Nejatian@nejatian·
I've literally never owned a car in my life. I'm buying a Tesla now *just* to give money to Elon.
English
96
33
967
83.7K
Alex Pilon
Alex Pilon@xaptronic·
@KatieMiller @kayleighmcenany I always thought Musk himself put it best when he referred to such wealth as "labour allocation" - namely when you successfully allocate labour to yield fruit, you receive more labour allocation to further your aims.
English
0
0
1
36
Katie Miller
Katie Miller@KatieMiller·
KAYLEIGH MCENANY: “I want to emphasize Elon Musk and what he has done for the world. “I stood alongside President Trump as SpaceX achieved America’s first crewed rocket launch in nearly a decade. It was SpaceX. “When Butch Wilmore and his crew were stranded, who saved them? It wasn’t the public sector — it was the private sector. It was @SpaceX. “@elonmusk’s companies — SpaceX, X, and Neuralink — are pushing humanity forward. “@neuralink is already helping restore communication and mobility for people with paralysis. That is Elon Musk. That is innovation. “And with innovation comes profit, and with profit comes the creation of millionaires and opportunity inside companies. That’s how capitalism works. It is not the great evil Bernie Sanders wants you to believe it is.”
English
96
800
4.6K
87.4K
Alex Pilon retweetledi
Daniel Beauchamp
Daniel Beauchamp@pushmatrix·
@seekingprotein Quick was built and is maintained by two people! @xaptronic and myself. And the maintenance now is very part time. Thing runs smooth due to so few moving parts
English
0
1
1
131
Alex Pilon retweetledi
Daniel Beauchamp
Daniel Beauchamp@pushmatrix·
The wildest part? All of Quick runs on a single VM with a single cloud database. In a high-trust environment, all the complexity you'd normally engineer around simply disappears. This thing only works because its internal.
Daniel Beauchamp tweet media
English
8
6
242
32.3K
Alex Pilon retweetledi
Daniel Beauchamp
Daniel Beauchamp@pushmatrix·
Everyone's talking about AI-generated HTML. But have you tried giving your sites a zero-config API for saving data, file storage, AI, websockets, etc? We did this at Shopify. Runs on a single VM that costs $200/month, and it's changed the way we work. We call it Quick 👇🧵
English
113
167
2.9K
835.2K
Alex Pilon
Alex Pilon@xaptronic·
@yacineMTB Can I be a programmer with no degree and not do this 😛
English
0
0
0
1.1K
kache
kache@yacineMTB·
hot take: if you are a programmer you should be able to invert a binary tree from memory, AI or not. It's ridiculously easy and if you can't do it, you should not have a computer science degree
Yuchen Jin@Yuchenj_UW

I’m so glad AI killed LeetCode interviews. For 10 years, tech companies made every engineer grind the same puzzles and prove they could invert a binary tree from memory. Today, the dumbest AI model can walk in and one-shot the entire interview. Thank you, AI.

English
287
64
2.2K
411.9K
Martin Mora
Martin Mora@Martinsmora3·
@pmarca @tobi Founders are great dictators (which they can be within their business), politics is a completely different game. You need to negotiate and end up only ever half solving problems.
English
2
0
6
2.5K
Alex Pilon
Alex Pilon@xaptronic·
@VivekGRamaswamy Lol blllllaaaahhhhh blah - there's been basically two political platforms. The people with scissors and the people with tape. Imagine what it would be like if there was something actually novel that could be said.
English
0
0
0
76
Vivek Ramaswamy
Vivek Ramaswamy@VivekGRamaswamy·
Ohio has over 2,000 taxing authorities, that’s way too many layers of government. We’re going to shrink bureaucracy, cut red tape, and put that $$ back in your pockets.
English
351
1.2K
12.6K
215.8K
Juno News
Juno News@junonewscom·
A reporter asks how the federal government is funding the $25B "Canada Strong Fund" endowment. Finance Minister François-Philippe Champagne points to the country's AAA credit rating and ability to borrow at low rates on international markets.
English
770
281
918
146.3K
Shaun McQuaker
Shaun McQuaker@shaunmcquaker·
Sachin Puri ... @bluehost is terrible. Every week there's an outage. Refuses to allow OpenAI crawl hosted websites. Customer service is clueless. You going to fix it or do I need to continue the public shaming for several more weeks for you to get your ducks in a row and fix it?
English
3
0
2
191
Alex Pilon
Alex Pilon@xaptronic·
@LeverCRO @Shopify @harleyf I found how many of the AI message founder responders are selling candles from their garage on Shopify 🤔
English
0
0
0
38
himanshu
himanshu@himanshustwts·
@muratcan too lazy for an X article, too fast for a scrappy tweet.
English
2
0
71
25.1K
himanshu
himanshu@himanshustwts·
Based on everything explored in the source code, here's the full technical recipe behind Claude Code's memory architecture: [shared by claude code] Claude Code’s memory system is actually insanely well-designed. It isn't like “store everything” but constrained, structured and self-healing memory. The architecture is doing a few very non-obvious things: > Memory = index, not storage + MEMORY.md is always loaded, but it’s just pointers (~150 chars/line) + actual knowledge lives outside, fetched only when needed > 3-layer design (bandwidth aware) + index (always) + topic files (on-demand) + transcripts (never read, only grep’d) > Strict write discipline + write to file → then update index + never dump content into the index + prevents entropy / context pollution > Background “memory rewriting” (autoDream) + merges, dedupes, removes contradictions + converts vague → absolute + aggressively prunes + memory is continuously edited, not appended > Staleness is first-class + if memory ≠ reality → memory is wrong + code-derived facts are never stored + index is forcibly truncated > Isolation matters + consolidation runs in a forked subagent + limited tools → prevents corruption of main context > Retrieval is skeptical, not blind + memory is a hint, not truth + model must verify before using > What they don’t store is the real insight + no debugging logs, no code structure, no PR history + if it’s derivable, don’t persist it
himanshu tweet media
English
151
692
6.3K
834.8K
Alex Pilon retweetledi
Daniel Beauchamp
Daniel Beauchamp@pushmatrix·
Huh, so that's why text is called a string
English
420
1.6K
20.6K
1.2M
Alex Pilon
Alex Pilon@xaptronic·
@garfield628 @harleyf @manofbird Worst take ever: technical innovation results in creation of more opportunity surface area than "drudge work" that it solves for. There's never been more possibility and opportunity for a larger number of people to tackle enormous unsolved problems.
English
0
0
1
15
mini vampire squid
mini vampire squid@garfield628·
@harleyf @manofbird the tech dudes only focus one-side view yet ignoring the fact that young people don’t believe in capitalism because the system doesn’t work in their favor. Look at how AI has diminished their prospects significantly.
English
3
0
0
313
Ted Bird 🇮🇱
Ted Bird 🇮🇱@manofbird·
The NDP leadership convention - an unintentional mockumentary at the level of Spinal Tap - concludes with a party that hates old white guys and Jews electing an old Jewish white guy as their leader. On a scale of 1 to 10, the irony goes to 11.
Ted Bird 🇮🇱 tweet media
English
117
305
1.8K
42.1K
Alex Pilon retweetledi
Eric S. Raymond
Eric S. Raymond@esrtweet·
I am a rather successful writer, and I never touch Grammarly, because my grasp of correct usage and my judgment of when to break the rules are better than its. I haven't heard of these other tools but I would be quite, quite surprised if they could exceed my competence.
Malvern, PA 🇺🇸 English
9
1
89
2.8K
Collings MacCrae 🧩
Collings MacCrae 🧩@CollingsMaccrae·
Listen. I’ve said it before, but here goes. Virtually all good writers use Ai as a tool. It’s in your auto editors such as Grammarly, PWA, AutoCrit. If you’re not using some tools, I’d say you’re working hard and not smart. I know some writers have blocked me for saying it: C’est la vie. You do you. However, this #ShyGirl kerfuffle is not the same. I have lots of questions. Lots.
GIF
English
56
3
95
10K
Andrej Karpathy
Andrej Karpathy@karpathy·
I'm not very happy with the code quality and I think agents bloat abstractions, have poor code aesthetics, are very prone to copy pasting code blocks and it's a mess, but at this point I stopped fighting it too hard and just moved on. The agents do not listen to my instructions in the AGENTS.md files. E.g. just as one example, no matter how many times I say something like: "Every line of code should do exactly one thing and use intermediate variables as a form of documentation" They will still "multitask" and create complex constructs where one line of code calls 2 functions and then indexes an array with the result. I think in principle I could use hooks or slash commands to clean this up but at some point just a shrug is easier. Yes I think LLM as a judge for soft rewards is in principle and long term slightly problematic (due to goodharting concerns), but in practice and for now I don't think we've picked the low hanging fruit yet here.
English
250
322
4.3K
821.7K
Andrej Karpathy
Andrej Karpathy@karpathy·
Thank you Sarah, my pleasure to come on the pod! And happy to do some more Q&A in the replies.
sarah guo@saranormous

Caught up with @karpathy for a new @NoPriorsPod: on the phase shift in engineering, AI psychosis, claws, AutoResearch, the opportunity for a SETI-at-Home like movement in AI, the model landscape, and second order effects 02:55 - What Capability Limits Remain? 06:15 - What Mastery of Coding Agents Looks Like 11:16 - Second Order Effects of Coding Agents 15:51 - Why AutoResearch 22:45 - Relevant Skills in the AI Era 28:25 - Model Speciation 32:30 - Collaboration Surfaces for Humans and AI 37:28 - Analysis of Jobs Market Data 48:25 - Open vs. Closed Source Models 53:51 - Autonomous Robotics and Atoms 1:00:59 - MicroGPT and Agentic Education 1:05:40 - End Thoughts

English
317
392
5.5K
1.2M
Alex Pilon
Alex Pilon@xaptronic·
@GergelyOrosz MCPs are opposite of dead because large slow moving companies use them? Skills+CLI replaces that complexity of maintaining additionally running services. Extract knowledge and skill from function.
English
0
0
0
20
Gergely Orosz
Gergely Orosz@GergelyOrosz·
MCPs are the opposite of dead. They are the life blood of how AI agents use services inside mid-sized and above companies. Case in point: Uber runs on MCPs internally, for good reason. Details: newsletter.pragmaticengineer.com/p/how-uber-use…
@levelsio@levelsio

Thank god MCP is dead Just as useless of an idea as LLMs.txt was It's all dumb abstractions that AI doesn't need because AI's are as smart as humans so they can just use what was already there which is APIs

English
92
95
1.3K
266.4K
Josef Strzibny
Josef Strzibny@strzibnyj·
Which project management software is more profitable? 1. Basecamp 2. Attlassian
English
16
2
47
12.4K
Kaz Nejatian
Kaz Nejatian@nejatian·
I am among the most AI-pilled human beings on the planet. Please don't try to build a CRM from scratch just because you can. Just use Hubspot. Or if you must, use something else off the shelf. Don't build. Buy.
English
128
55
1.3K
136.8K