José Luis (@ACJluis) header
José Luis (@ACJluis) profile picture
Joined September 2012
Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)Photo by José Luis (@ACJluis)
José Luis retweeted
the only moat left is the love of the game
the only moat left is the love of the game
1
3
15
863
José Luis retweeted
the moat is literally just caring about your work and being proud of your output and holy shit it does not seem to apply to 99% of people
108
512
69
7,470
382,498
José Luis retweeted
When you do boring shit, you tend to over engineer it When you do ambitious shit, you’re forced to keep it simple Great lessons in there
38
347
22
4,721
76,201
José Luis retweeted
adulthood is realizing your dream job isn’t separable from the social world it puts you in
14
53
4
1,030
41,080
José Luis retweeted
An alternative hypothesis: - model performance is plateauing - compute is getting much more expensive - AI data centers are massively unpopular - slowing down AI development is a way to explain slowing progress, reduce spending, and try to regain some goodwill. - This is an effort to save the IPO not humanity.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: darioamodei.com/post/we-must…
269
2,337
212
19,990
915,949
José Luis retweeted
my best advice for young guys going into the trades is to make $250-500k in an email job, then buy a bunch of tools and a huge garage and start a custom knife business or a bespoke woodworking shop where you lose $100 on every sale
114
393
30
16,157
334,925
this meeting could have been an MCP invoked search of our company’s internal knowledge base that aggregates data from virtually every service we use behind a unified retrieval layer and provides a streamlined interface for extracting critical business insights
94
547
37
6,849
270,757
you, an imbecile, learning to prompt LLMs better me, also an imbecile, learning to prompt google better
117
1,163
103
33,731
1,691,802
José Luis retweeted
You know how feeling like you lack the energy to exercise is a good indicator you need to exercise? I feel like a similar thing applies but for cultivating a zest for life; feeling bored by artistic pursuits, hobbies, etc is a sign you need to develop them
9
779
18
9,226
106,672
José Luis retweeted
I like how noticeably “fake” facial expressions are often actually completely honest, but are just carrying a more specific message. If the goal were deception, it is well within most people’s capacity to fake an actually convincing smile. However, fake smiles communicate “I’m just being polite and am not actually smiling.” And you’re actually meant to notice all of the information conveyed and behave accordingly.
14
8
2
417
18,132
José Luis retweeted
behind every published book is a typo that made it through 1.5 million rounds of edits
46
146
68
2,038
3,020,027
Dostoevsky was right; “Every self-betrayal is a sin. Whenever you go against your nature, your body reminds you.” If you spend enough time with anything, you start liking it, even sadness. So let’s choose people and spaces that truly elevate us. Your peace is worth it.
JUST IN🚨: According to Research, the more we focus on beautiful things, the more the brain gets used to seeing beautiful things.
23
3,128
33
19,369
428,240
Replying to @BlindarMH
@BlindarMH @cesar_m_garrido @PAOTmx nos pueden apoyar con esto en la TORRE KLU montes urales 405 el ruido a todo @pilfernandez @Copalomas @comitebarrilaco
5
8
1
6
439
José Luis retweeted
yo el 30 de junio después de que claudia sheinbaum me suspenda la línea telefónica
56
10,014
148
52,758
629,396
José Luis retweeted
Just remember: 68% men have visible abs 86% men bench 220 lbs 90% men run sub 20 5k 113% men squat 315 lbs You're doing worse than you thought
572
1,631
177
59,782
4,522,806
José Luis retweeted
fiber doesn't "fill you up" soluble fiber is fermented by gut bacteria into short-chain fatty acids. those SCFAs stimulate L-cells in your intestine to release GLP-1 (the same hormone that ozempic mimics) you can make your own GLP-1s, for free, from fiber
34
28
7
1,082
43,865
José Luis retweeted
My man said something to me that really stuck. He told me, “I’m not here to control you. I’m not your dad, I’m your partner. You’re free to make your own choices. Just understand that every choice has consequences. If you choose something that damages what we’ve built, that’s on you.” He said, “I’ll always tell you when something hurts me or crosses a boundary, because that’s what healthy communication looks like. But if you keep stepping over the line after I’ve shown you where it is, then you were never really protecting us to begin with.” And honestly, that’s what accountability in a relationship sounds like.
Unpopular opinion about marriage that would get you in this position???
234
4,821
150
28,628
2,417,063
José Luis retweeted
The moment you threaten to leave a man or end a relationship, make sure you say it once and back it up. Cos you’re never getting back the same man that once loved you.
Men, give women advice from a man’s POV.
134
3,780
100
22,805
1,075,901
José Luis retweeted
LLM Knowledge Bases Something I'm finding very useful recently: using LLMs to build personal knowledge bases for various topics of research interest. In this way, a large fraction of my recent token throughput is going less into manipulating code, and more into manipulating knowledge (stored as markdown and images). The latest LLMs are quite good at it. So: Data ingest: I index source documents (articles, papers, repos, datasets, images, etc.) into a raw/ directory, then I use an LLM to incrementally "compile" a wiki, which is just a collection of .md files in a directory structure. The wiki includes summaries of all the data in raw/, backlinks, and then it categorizes data into concepts, writes articles for them, and links them all. To convert web articles into .md files I like to use the Obsidian Web Clipper extension, and then I also use a hotkey to download all the related images to local so that my LLM can easily reference them. IDE: I use Obsidian as the IDE "frontend" where I can view the raw data, the the compiled wiki, and the derived visualizations. Important to note that the LLM writes and maintains all of the data of the wiki, I rarely touch it directly. I've played with a few Obsidian plugins to render and view data in other ways (e.g. Marp for slides). Q&A: Where things get interesting is that once your wiki is big enough (e.g. mine on some recent research is ~100 articles and ~400K words), you can ask your LLM agent all kinds of complex questions against the wiki, and it will go off, research the answers, etc. I thought I had to reach for fancy RAG, but the LLM has been pretty good about auto-maintaining index files and brief summaries of all the documents and it reads all the important related data fairly easily at this ~small scale. Output: Instead of getting answers in text/terminal, I like to have it render markdown files for me, or slide shows (Marp format), or matplotlib images, all of which I then view again in Obsidian. You can imagine many other visual output formats depending on the query. Often, I end up "filing" the outputs back into the wiki to enhance it for further queries. So my own explorations and queries always "add up" in the knowledge base. Linting: I've run some LLM "health checks" over the wiki to e.g. find inconsistent data, impute missing data (with web searchers), find interesting connections for new article candidates, etc., to incrementally clean up the wiki and enhance its overall data integrity. The LLMs are quite good at suggesting further questions to ask and look into. Extra tools: I find myself developing additional tools to process the data, e.g. I vibe coded a small and naive search engine over the wiki, which I both use directly (in a web ui), but more often I want to hand it off to an LLM via CLI as a tool for larger queries. Further explorations: As the repo grows, the natural desire is to also think about synthetic data generation + finetuning to have your LLM "know" the data in its weights instead of just context windows. TLDR: raw data from a given number of sources is collected, then compiled by an LLM into a .md wiki, then operated on by various CLIs by the LLM to do Q&A and to incrementally enhance the wiki, and all of it viewable in Obsidian. You rarely ever write or edit the wiki manually, it's the domain of the LLM. I think there is room here for an incredible new product instead of a hacky collection of scripts.
2,938
7,399
2,223
61,104
22,038,817
José Luis retweeted
If you wait too long, the coffee gets cold, the door closes, you get old, the girls move on, and dreams fade. You must act with a sense of urgency today.
Mature men who are 25+, please drop one piece of advice to men 18 to 24. It can be about anything!
44
1,459
11
14,457
453,015