hungry hungry hippos
I saw a video that said @sama is trying to abolish capitalism entirely. Through the means of increasingly cheap intellectual work (AI), eventually, everything will be worth nothing making capital meaningless.
I would argue he's the ultimate capitalist because OpenAI's goal is to capture all means of production leading to a collapse of meaningful wealth and value outside of a winner take all AI monopoly.
Sam, which is it?
94
RSPS community shines again (the other side of the coin to the economic education that playing rs as a player gives you)
Replying to @WilliamHolmbe19
This vibe-coding thing? It’s in my blood.
At 9, I’m running RuneScape private servers—hacking together janky worlds, tweaking drops, spawning chaos with friends.
By 10, I’m coding bots—scripting them to grind 24/7. At 17, I’ve got a bot-net: 10 machines, 300 bots, running like a goddamn empire. Didn’t care if the code was trash—I’d ship it, break it, fix it, ship it again. Pure joy, no overthinking.
@levelsio style? That’s my childhood in a nutshell—just build, fuck perfection.
Then I hit software jobs. Shit changed. Suddenly it’s all “best practices,” “code reviews,” “scalable architecture.”
I get it—quality matters. But over years, that “just ship” fire died.
I’d sit there paralyzed, tweaking shit endlessly, scared it wasn’t “good enough.” The kid who’d slap a bot together in a night and laugh when it crashed? Gone.
Work beat the joy out of me, and I didn’t even notice until Glenn brought it back.
1
1
135
old school runescape teaches typing past 120wpm and lightning fast spatial awareness as a survival skill
Giving very young kids access to real laptops with high ceiling video games is such a cheat code. My kids bootstrapped reading, math, spatial reasoning, etc, just because they had a *need* to do it since the carefully chosen games I've put them in front of *demand* it of them.
2
138
seniors should not age in isolation
Replying to @Thrilluwu
This isn't surprising at all. VR used in any situation where isolation is an issue will be a huge impact. There needs to be more attention brought to seniors with VR. There is no reason anyone should be aging in isolation with the technology we have.
59
“doing it in time”
there are always at least 6 other teams of people around the world with access to the same information you have, and have reached the same conclusions you have.
the timer is always ticking. execute, or be beaten to the finish line.
Replying to @corry_wang
7/ In the end, Nokia didn't launch its first "real" iPhone competitor (the Lumia 800) until 4.5 years after this deck... and the resolution of the Symbian / Maemo struggle
The hard part about the innovator's dilemma is not knowing what to do, but actually doing it in time
69
Mahdi Chaker retweeted
I've been telling you for a while not even to try to compete with machines at being a better machine. Instead, try competing with humans at being a better human.
Mahdi Chaker retweeted
I’m increasingly coming around to the idea that the natural state is joy
other emotions may arise, but if left alone and felt fully, they pass and transmute, leaving joy behind
and it seems that joy is often there as a background tone even if other emotions are flowing through?
Mahdi Chaker retweeted
★ Dynamicland's new website documents ten years of progress toward a humane dynamic medium. dynamicland.org
Good way to understand ADHD.
Replying to @scottastevenson
My post might make it sound like large context windows are obviously better—but that’s not true.
Relying on long context windows often means oversights, forgetfulness, hallucination. Whereas RAG and other breakdown strategies can actually increase accuracy.
Personally I do feel firmly in the second camp though. When I designed I’d always go straight to CSS. When I coded I would always go straight to writing code. When doing marketing I’d go straight to building the landing page.
Many (including teachers) have tried to intervene and show me the light in more structured ways of working, but my brain just can’t grasp breakdown strategies a lot of the time.
1
60
Mahdi Chaker retweeted
After working with LLMs for 2 years, I’ve noticed that humans have very different short-term context windows too.
Humans with small-to-mid sized context windows develop deep “breakdown” strategies early in life to deal with virtually everything. Ways of breaking down complex problems into smaller steps that can be chained. Their brains contain an enormous recipe book of these multistep breakdown strategies.
Humans with very large context windows operate off raw intuition for a lot of life. They did not develop many of these multi-step prompting strategies as a child, and the idea of “processifying” work or breaking it down into smaller steps is *extremely* puzzling—feeling like it will only slow them down and confuse things.
Both types of people have their strengths and weaknesses, neither is “better”.
But man, we do not understand eachother, and how different our ways of interacting with the world are.
Mahdi Chaker retweeted
This is a great little book to take you from “vaguely understand neural networks” to the modern broad state of practice. I saw very little to quibble with. fleuret.org/francois/lbdl.ht…
A worthy leader loves and encourages and trusts his reports *in the way they know how to operate* — meet them where they are. I am glad to have experienced this and want to hold on to it.
Another point for the "gentle self-love isn't all you need" camp
I've had a really vengeful warrior part of my personality for over a decade now. He's been a great source of strength and self-discipline, but also troublesome anger and impulsivity. He constantly hijacked my attention to brood over threats and rehearse potential conflicts. I tried everything I could think of, but nothing got him fully under control.
Then one day, I realized: my true self has the divine right to rule. Like, literally has the claim to the throne of attention and presence. So I centered, brought out the warrior part, and like a king, demanded he swear undying fealty to the self called "I am."
Surprisingly, the warrior part was overcome with relief, and immediately submitted. He had been serving my disowned desire for revenge - a god that didn't love him - and was actually desperate for a worthy leader.
So I retired him from active duty and gave him a job as chief of staff. Now he's in charge of structure, discipline and boundaries, but he takes orders from me. Mindful presence has been mostly effortless for me since then, and joy, laughter and safety have come much easier. The warrior's habit energy is still there, but obvious to me now. We have mutual trust that I'll deploy him when he's needed, and he'll fight like hell for me when he's called on.
The moral of the story is, you can try many different orientations and exchanges with your multitudes, and see what yields results. Gentle self-love is excellent as an all-purpose tool, but sometimes specialized parts need specialized relationships, according to their internal value and logic structure. Meeting them where they're at is its own kind of compassion.
105
Mahdi Chaker retweeted
“The present is the past rolled up for action. The past is the present unrolled for understanding.”
I too plant this flag for the kind of love described in this thread: total surrender, progressively deepening love radiating around me everywhere. Appreciate. Appreciate. Appreciate. Keep it simple :)
Replying to @newtonlaw
It does seem that we’re entering times of great upheaval. I suspect we’ll need this kind of love more than ever. I’m planting a flag to say that I’m here for that love.
2
113
Mahdi Chaker retweeted
Understanding that every compressor is a machine learning predictor, and vice versa, was the single most important insight I learnt about between 2019 and now.
this is wild — kNN using a gzip-based distance metric outperforms BERT and other neural methods for OOD sentence classification
intuition: 2 texts similar if cat-ing one to the other barely increases gzip size
no training, no tuning, no params — this is the entire algorithm:
![for (x1, _) in test_set: Cx1 = len(gzip.compress(x1.encode())) distance_from_x1 = [] for (x2, _) in training_set: Cx2 = len(gzip.compress(x2. encode()) x1x2 = " ".join([x1, x2]) Cx1x2 = len(gzip.compress(x1x2. encode()) ncd = (Cx1x2 - min(Cx1,Cx2)) / max(Cx1, Cx2) distance_from_x1.append(ncd) sorted_idx = np.argsort(np.array(distance_from_x1)) top_k_class = training_set[sorted_idx[:k], 1] predict_class = max(set(top_k_class), key=top_k_class.count)](https://pbs.twimg.com/media/F05HxqsWwAAHRCI.jpg?name=small&format=webp)
ALT for (x1, _) in test_set: Cx1 = len(gzip.compress(x1.encode())) distance_from_x1 = [] for (x2, _) in training_set: Cx2 = len(gzip.compress(x2. encode()) x1x2 = " ".join([x1, x2]) Cx1x2 = len(gzip.compress(x1x2. encode()) ncd = (Cx1x2 - min(Cx1,Cx2)) / max(Cx1, Cx2) distance_from_x1.append(ncd) sorted_idx = np.argsort(np.array(distance_from_x1)) top_k_class = training_set[sorted_idx[:k], 1] predict_class = max(set(top_k_class), key=top_k_class.count)
Mahdi Chaker retweeted
Texts with Founders: Avoiding Gossip
I recently had a founder reach out to me after someone pressured them to share private details about another person’s life. They didn’t want to share the information but didn’t want to create a tense situation.
When a person tries to get you to gossip about others, it can feel awkward to tell them it’s none of their business.
One way of refusing the person’s request for gossip in a non-confrontational way is to say the following:
“I don’t discuss people’s private matters with others. This way, if you ever wish to share something with me that you want kept between us, you know it will be.”
This response firmly rebukes the request without directly calling the person out for an inappropriate ask. It also shows them that you are a trustworthy person.
At the risk of stating the obvious, you should strongly consider whether it’s worth risking sharing any sensitive information with this person going forward. Your personal information might become future gossip for them to spread around.
innovation comes from play
Replying to @andrew__reed
doesn’t feel like people talk about why that doesn’t work:
it’s because he did it because he followed curiosity! it’s not about whether you can do it/replicate it now, it’s that you need to replicate how he got to that in the first place.
1
97





















