//Krumil

Web3 and AI developer.
Always looking to innovate.
Feel free to DM me for any ideas.

[//Krumil @Simo1028]

I'm excited to share a project I've been working on called Satoshi Sentients. This agent generator leverages core technology from @BrianknowsAI and @langchain, along with the @AnthropicAI Claude model, to make trading decisions.

[//Krumil @Simo1028]

@z @alex_frantic Symphny is super interesting and confirm the experience of many of us, but I think there's another step to be made in order to fully understand agentic coding. Something that give coders a more reliable way to keep codebase on track and checking progression at higher level

[//Krumil @Simo1028]

@CoinbaseSupport Any chance on sharing the evaluation framework? Would be very useful for what we are building

[//Krumil @Simo1028]

@RiftboundMeta @SouledOutSoon Exactly. Metal cards reward you for brewing without any obligation. Great idea

[//Krumil @Simo1028]

@Fiskantes https://t.co/RN2bVaJ9Gq

[//Krumil @Simo1028]

@cleervision @dabit3 Hey can I get more info on what you are building? I'm definitely fully pilled

[//Krumil @Simo1028]

@IamActuallyLvL1 It's an incredible idea imo

[//Krumil @Simo1028]

@Necrit94 It's just an entire meta being build against it. Now, we will see more ezreal/gear hate and it will shine again

[//Krumil @Simo1028]

@SiloFinance Hello, discord link doesn't work. How can I access it? Thanks

[blockgraze @blockgraze]

"bro have you tried clawdbot it's so insane"

"a little but not much, what do you use it for"

"it's crazy many you can do anything with it"

"what are you doing with it"

"you gotta try it"

"try it for what"

"don't get left behind man"

[//Krumil @Simo1028]

@LotharHS @riftboundop @Silv4nn Hi Lothar! Sent you a dm about your super useful sheet for the meta https://t.co/6y9Z81laJE

[//Krumil @Simo1028]

@LotharHS Hey man! I know you are at Regionals right now, but I found your sheet so useful that I made a site for it https://t.co/GwFNjsyo8q. Let me know if you think this could be interesting to mantain

[//Krumil @Simo1028]

@LotharHS Hey man! I know you are at Regionals right now, but I found your sheet so useful that I made a site for it https://t.co/GwFNjsyo8q. Let me know if you think this could be interesting to mantain

[//Krumil @Simo1028]

@0xDesigner Ai created my job

[//Krumil @Simo1028]

@ObserverJohnny @RiftboundMeta A seal gets the same result. Card is definitely strong. It lets you see 10% of your deck and draw 2 😅

[//Krumil @Simo1028]

@tengyanAI But honestly what's the alternative here? Laws need accountability and accountability is a human prerogative...

[//Krumil @Simo1028]

@TheRealAdamG Whoever has done that is lagging behind

[//Krumil @Simo1028]

@theo This is my number 1 concern right now. I've lost more time switching between terminals/tabs than sleeping

[//Krumil @Simo1028]

Yes pls! I think that we are still under using agents for brainstorming and ideas.

I'm building a project that will come up with new ideas for work/life/hobbies by its own. You select interests/topic, multiple agents works on that and ask you questions (or not) and cama up with a boilerplate implementation

[//Krumil @Simo1028]

@SorcererTrading Hi!

[vitalik.eth @VitalikButerin]

I'll do the ill-advised thing and try to explain my own thought process and constraints (and possibly unadmitted cowardice) that guide when I do and don't speak on these kinds of political topics - and further down just say what some of my direct opinions are. It's 2026, the careful route isn't getting us anywhere anymore, might as well try being open.

When it comes to hot political issues (not abstract questions like tax policy, surveillance, etc, which it's actually easy to talk about, I mean specific events that affect specific groups), I basically have three choices:

1. Talk about none of them
2. Talk about some of them
3. Talk about all of them

Many people do (1). In crypto, many people talk about "decentralized governance", a "freedom", a "fairer economy without intermediaries", "cypherpunk", etc etc, and maybe talk about how blockchains and encryption and zero knowledge proofs can make these things happen. But they take care to avoid making the link between those values and any specific situation. This approach is "safe". But ultimately, it feels hollow, and I think it makes your mind hollow. It makes it really easy for you to think you're doing the right thing, because you're working on all the right technologies, but then because you never engage on any concrete issue, you don't even notice if you're not actually making any impact on the underlying problems - or worse, actively diverting idealistic talent and effort from actually-effective solutions. I think this applies both to individuals, and to whole communities. And so if you don't want to fall into this trap, you have to engage with the world.

Note that this does not mean that *your work* can't be general-purpose and instead needs to be targeted to directly affect some specific situation in the world. Focused work is good. But I do think the world benefits from a decentralized public apparatus of moral conscience that extends beyond topics that you directly work on.

Many people do (3). But there is a good argument against (3): if you are forced to take positions on everything in the world, then you're usually taking positions on things you had little or no prior understanding of, and your positions are motivated by a few emotional articles or posts that recently convinced you in one direction or the other. And if you have 100 topics, you can only devote 1% of the effort to tracking each one. It becomes easy for someone to convince you of a position that you would not even support if you were more informed about it. There is a reason why people hate the "omnicause", "everything-bagel activists", etc. Focusing on everything dilutes the message to the point where you're succeeding at nothing (or worse, having actively counterproductive postures on things as a result of low information).

Hence, there is the option of (2). But (2) has a natural problem: once you talk about some things, it becomes easier for people to pressure you to talk about even more things, by accusing you of selective outrage and hypocrisy. (And, to be fair, selective outrage and hypocrisy are very real problems) To defend against this, you need to hold "a line". One very natural line is to mostly limit your attention to topics that relate to you personally.

For example, I and my family are from Russia. When I see footage of the war in Ukraine, whether videos of soldiers getting ready for a military operation, or families expressing their anguish after their apartment building or hospital was bombed, they are often speaking in a language I have known since childhood. More importantly, and regrettably, I personally met Vladimir Putin and thus to some small degree helped legitimize him back in 2018. To me, these things mean this conflict relates to me and is my responsibility to do the right thing in (and not just be a passive bystander, perhaps saying "I am against war" exactly once and then continuing my life).

Similarly, Canada is my responsibility, as it's the land where I grew up - both when the government financially deplatformed the truckers (which I criticized), and when the bumbling old man down south starts bullying and threatening its sovereignty (which I've also criticized).

Meanwhile I've spent 0 days in Myanmar or Venezuela, less than three months in the Middle East, and have much less context in those and other places - I know what I know from reading second-hand sources and making my own judgements about which facts are true and false and which arguments are right and wrong, which is far from zero, but it still has limits.

The United States is a special case, because historically politics in the USA affects the world so much, many people (though today much less than a decade ago) look up to it as an example, and it has massive influence through its economy and its centralized technology platforms. So sorry guys, the entire world has a right to blab about your internal affairs. Something something vaguely similar to "no taxation without representation".

This approach that I outlined does have a weakness: regions of the world that are economically very poor will have very few globally powerful people that have close connections to them. Hence, a "take care of your own" norm leaves such people in the dust, vulnerable to being ignored or even outright predated on by others, with no one powerful sticking up for them. I personally try to address this by first making a judgement about whether a faraway situation is more like an internal conflict or more like eg. global public health, and apply the "take care of your own" norm in cases like the former and not in cases like the latter.

This is the explanation I can give for why I've publicly said relatively more about Canada, USA and Russia, and relatively little about Venezuela, Sudan, Africa, Myanmar, China, or the Middle East, including both Iran and the various conflicts involving Israel. Basically, if I don't draw the line roughly here, what other line can I hold?

*Maybe* there are some people who are closer to "pure devs" who can hold a different line, speaking *only* about abstract generalities, but I've always covered the full spectrum from pure tech to social issues, so I don't think I *can* hold that line. Hence, instead I have to hold slome further-out and more complicated line.

Of course, it's also easy to construct a less self-serving explanation: I have some view of who my constituency is, and I am a coward who is afraid of offending them. I could respond by pointing to various instances in which I've been publicly brave (whether calling out Craig Wright, or visiting Kyiv, or running through 2km of rain to get to a conference panel on time etc), but maybe my critics can justly counter by pointing out that this is simply the old-as-time masculine tactic of "compensating for something".

I will let readers make their own judgements.

Now, my views on a few specific topics:

Iran

From what I can tell, the Iranian regime is:

* Literally killing tens of thousands of people, in gruesome ways, in its crackdown on the current protests
* Totally denying Iranian people access to the worldwide internet, in part to cover up the above, and there is a high chance that the regime intends this situation to be permanent.
* In the longer term, doing more generally awful things like imposing what clothes women have to wear, providing military support to Russia's invasion, publicly wishing death to various groups on official channels, etc

These things are unambiguously awful. They're not a "eh, what can we do" normal level of awful, they are totally awful. Even if global public condemnation accomplishes nothing else, simply reducing the regime's social status to the point where it equals that of North Korea's gov seems like part of the bare minimum that should be done. Though hopefully, the protests will succeed and expand and Iranian people will get freedom. I hope that the crypto space finds a way to be useful by exploring more options to restore access to the global internet to Iranians.

Furthermore, there is another important point to make: many people say "this is awful", but then still maintain an attitude of disapproval toward overly "impolite" ways of dealing with the problem, eg. using physical violence against the Iranian leadership. I see the value of a "suggesting violence is impolite" norm in maintaining civilization, but IMO that norm should be focused on *initiating* the escalation to full-scale violence; it's fine to respond with violence when violence is already being used in the other direction on a large scale. Being overly pacifist is a good way to feel self-righteous while doing nothing effective to prevent the people you sympathize with from being rolled over.

Nothing in those above paragraphs is "anti-Islamic". But there is the question: if we analyze the situation, and ask *what causes* the Iranian government to be abnormally brutal, then might the conclusion be that those causes include cultural values that are core to Islam?

I personally don't understand Islam enough to comment. But I can say two things:

* There's plenty of extreme evil that happens without Islam. the invasions carried out by Russia are one good example, there's plenty of extremely dangerous political megalomania worldwide that is 100% secular
* In my experience, the anti-Islam hate that I see in our world often goes beyond cultural criticism, and turns into attacks on people and condemnation of entire ethnic groups (see eg. post 9/11).

Culture can and should be criticized, including by outsiders. This includes banal things, like my view that miles, feet and pounds are stupid units of measurement and SI units are more civilized (I have much less confidence on 'F vs 'C). It also includes deeper things like moral values. But there is a way to do this that does not turn into personal attacks on people. I do often see anti-Islam discourse failing at this (though some criticize effectively and humanely).

Israel/Gaza

My moral view on the situation is:

* The attack on Oct 7 was a brutal unjustified war crime that killed around a thousand innocent people, and should be condemned without qualifications
* The response to the attack has killed ~50-100x more people than the original attack, the majority of them civilians, and has destroyed the homes of over a million Gazans and traumatized the entire population. This level of total devastation in response to an attack of a much smaller scale exceeds all reasonable bounds of self-defense, and itself is a brutal war crime
* There are various statements made by high-level officials, actions by soldiers, etc, that I've seen that suggest that the mental state involved in the Gaza operation is less "a principled effort to protect people" and more "classic pre-modern revenge psychology"
* Hence, the ICC was right to charge both sides of this
* The Israeli government is on the whole a lot more rights-respecting and better to live under than the Hamas one. However, the Hamas government already has social status worse than North Korea. Meanwhile, the Israeli one often gets treated at the highest official levels as a special valued ally, in a way that I fear deeply undermines moral norms, and thus creates room for eg. letting Putin get away with his crimes in Ukraine. This is an imbalance, and I think this is something that many people are reacting against.
* What happened in 1967, 1948, and arguably even the late 2000s is honestly not very relevant to this moral analysis. The majority of the population was not even alive when those things happened, there's no good moral reason why the fate of a 14 year old kid who is seeing their home or family destroyed should depend on arcane debates about the previous century's history
* The biggest heroes in this situation are probably people quietly working in the background to reconnect people and lay the ground for peace that will hopefully shape itself over decades. I don't think the world can respond *only* by doing this, because it's also important to defend norms ("injustice anywhere is a threat to justice anywhere"), but I'm really glad these people exist and morally support them.

I strongly oppose anti-Semitism. To me, this is a special case of the broader moral principle that one should not judge entire ethnic groups by the decisions of a few elites. In other worlds, "hate the government, not the people". One important corollary of this, of course, is that criticizing the Israeli government does not count as anti-Semitism.

Principles vs details of the situation

I should note explicitly that these moral views above are very "principles-driven" and not "details of the situation" driven. I think this is the correct level at which to approach this. The reason is that while "details of the situation" thinking often can acknowledge subtle important facts that would otherwise be ignored, nevertheless:

* It has very low galaxy brain resistance ( https://t.co/iJYPFNagqI ). It's too easy to cherry-pick details that favor your side, and appeal to particularism ("you outsiders can't understand...") to shield one's local bullying from external pushback
* It's not something you can build norms around. Peace and humanitarianism in the world depend on moral norms: the fact that the whole world recoils in disgust at certain unconscionable outcomes, and that it can do this in lockstep, and can't be divided-and-conquered by arguing arcane points about third-order game theory or history. Some say "norms are fake anyway, we're in the law of the jungle, you have to protect your own". To anyone who sincerely believes this, all I can say is: YOU HAVE NO FRIGGIN IDEA HOW MUCH EVERYONE, EVEN THE VILLAINS, ARE HOLDING THEMSELVES BACK BECAUSE OF MORALITY AND NORMS EVEN TODAY, AND HENCE, HOW MUCH WORSE THINGS COULD GET
* Outsiders to a situation have more understanding of principles than of situational details. Hence, if it's valid for outsiders to express views on things at all, they should focus on their relative expertise, which is the principles.

ICE and Immigration

The other big thing happening now, that I've been silent on, is ICE turning into a full-on police state and now shooting protesters in broad daylight.

The immigration issue is complex and it is important to separate two aspects:

* Whether less or more immigration is good (and what kinds). If the highest acceptable level of immigration is less than the level that would naturally arise in the absence of either walls or deportation, then walls or deportation are in principle required.
* The fact that ICE is acting like total assholes about this situation.

These things can and must be separated. There are immigration systems far more restrictive than USA, whose implementation is more humane.

The second needs to be strongly opposed, full stop, no counterarguments admitted. Once the police state apparatus exists to this extent, it *will* keep finding new targets. Even today, it has already expanded its violence from illegal immigrants to obviously-American defenders of immigrants. [insert Niemöller poem here]

Today, this second issue (the cruelty) is the more important one in the USA, and its morality is pretty cut-and-dried simple, as described above, so nothing more to say about it. So now, on to the first issue (ideal levels of immigration).

I am generally persuaded by Bryan Caplan-style arguments that the answer is "very high". To summarize:

(i) the bulk of economic evidence suggests that high-skilled immigration is pretty close to no-downsides good, high-volume low-skilled immigration *slightly* depresses wages of low-wage locals, but it also means consumer prices go down for everyone. The $3 falafels in Berlin (which in turn support the city's status as an affordable home for various artistic and cultural activities) are only possible because of the Middle Eastern low-skill labor
(ii) in the USA, even illegal immigrants (!!!) on average commit less crime than the native population. In Europe though this is not true.

I also add my own bespoke point (iii): if you are worried about change to *culture*, then I believe that the larger driver of culture change is not foreigners but technology. The real great replacement is the kids growing up on TikTok. (also, much of cross-country culture spread happens through the internet, not via farm workers or Uber drivers)

And so actually I believe that if you care about stopping the great replacement, the right thing to invest in is ... longevity technology, so that existing human beings and their generations with unique cultures can survive through the ages as human beings, and not just as history pages. That's right: making everyone live 10x longer is a *conservative* technology.

(To the counterargument that this will also make evil dictators live 10x longer and prevent their societies from ever escaping: if we get into that world and a dictator gets too awful, then yeah I'm ok with droning them, I already said so above)

However, the downsides to public safety and government welfare expenditure from some types of immigrants are real. Personally, I think that we're very far from exhausting the opportunities we have to solve this problem without deporting people at all - I regularly see stories of criminals getting freed after committing three murders and then quickly going out and committing a fourth. We should stop doing this. Every act of lenience against a violent criminal saps public acceptance for immigration, and thus leads to ten acts of anti-lenience toward peaceful foreigners. Hence, true long-term empathy would withhold lenience from the criminals, and apply that lenience to peaceful foreigners instead.

But there is a larger problem behind this problem. Western morality is very dominated by "out of sight out of mind" bias. If you are on the inside of the wall, then you're part of the family, all human beings are equal, let's sing and dance in a circle. If you are on the outside of the wall, then you get performative thoughts and prayers, but in terms of substantive support, well, hang on we'll get right back to you.

This is where I think the social role of the "illegal immigrant" category in the USA comes in. It is a strange category, because we're saying that legally there are tens of millions of people who are violating the law on a mass scale, but we're not _doing_ that much about it. Requiring employers to verify immigration status of employees frequently gets shot down. But from the perspective of "Western society hacking around its own broken moral code", it makes perfect sense. We believe in "all humans are equal", that "second-class citizenship" is a bad and dystopian idea, that everyone deserves "good" working conditions, etc etc. But the economic reality is that the wage level that maximizes (not only local economic, but also humanitarian) gains from hiring foreign labor is lower than the level that is acceptable to locals - if you force the wage for low-skill foreigners up to the same level as locals, then you end up hiring much fewer of them, and so even the total humanitarian benefit you bring to low-income foreigners by giving them more economic opportunity is lower. So how do we solve this? Well, we invent the "illegal immigrant" category, and so the moral baseline becomes "well technically you don't even have the right to be here, and so really, anything you get here counts as an extra bonus". This is the moral self-arbitrage.

The problem is that wealthy countries by default are exactly in the position of the space colony from the movie Elysium: they're gated communities for the global rich, where everyone outside is a second-class citizen on a global scale. So if we say "we refuse to exploit foreign labor, they can all stay outside", we're not actually being moral.

Hence, to me the principled solution seems to be, to be more willing as a society to say "you're welcome in the land and you can work, but you're not part of the family", and to create visa categories that reflect this fact. Paths to become part of the family should continue to exist, but it's okay for those paths to be much more difficult than paths to simply come in and work. This is a model that already works well in many places in the non-West. And today, I find many non-Western countries increasingly opening themselves up to foreigners, while Western countries close down, add more ESTA and visa bureaucracy, border guards become more annoying and insufferable, etc.

The grand sacrifice: take less seriously the idea that "once you're within the walls, everyone is part of the family and is equal". This mentality was more sustainable in the 20th century, when people's identities were mostly bound up with a single country and most people did not even see anything outside their home country, but it is incompatible with the 21st, where information is global and our identities are plural. Unless we can achieve equality across the whole world, saying "within our home we are equal, outside our home go away" is ultimately a larp morality, and we should move on to something better. This *should* be a major task of the next generation of compassion-driven political philosophers. If we do this, it opens the door to greater openness to the outside world, which will ultimately be a human rights boon to everyone by giving people much-needed second options.

(special thanks to @ameensol for pushing me to be more brave even though I predict he will disagree with ~1/3 of the above)