holmesworcester

One way to settle the question of whether Khan is right would be for the government to simply make competing offers in these situations, buy the companies, and shepherd them to IPO, or a buyer with fewer antitrust issues if that's not possible.

If the government is net ahead after a decade or so, then we'd know.

This approach to antitrust wouldn't work in cases like the Apple case, where the power is worth it to the company only because they can misuse it, but it would be a very fair and accounting-transparent remedy for the "startup gets bought by competitor" case.

Then why did shareholders choose to sell?

In choosing to sell they decided the risk wasn't worth the reward.

If you were in their position, would you have sold or held?

I'm sympathetic to a prohibition on big companies buying their competitors, but a 3x difference over two years seems too low to suggest that antitrust creates more pure business value.

First this is all hindsight now. We don't know the probabilities of this outcome vs. others. Figma's shareholders didn't at the time, which is why they chose to sell. Khan didn't either.

Second, 3x over two years isn't that much. There must be many opportunities in SV for all of Figma's employees and investors that could have given them a much higher return than that with much less risk.

I don't have this data, but one could look at secondary sales in the past two years as a measure of the increased risk and opportunity cost, right?

Any delay of people getting liquid impacts the creation of other startups, both by the Figma people who can now leave and do their own thing and for the companies Figma stakeholders would have invested in . This is super hard to measure but it is the kind of thing markets are good at measuring when they ask shareholders "sell now to Adobe or wait to IPO?"

This seems really good for Figma users, most of all. Most of the value destroyed by the acquisition would have been in the distortion and likely ultimate destruction of a company culture that made an insanely good product.

But those people are capable of going and making new products, and maybe Figma at its current phase is now too boring a thing for their talents, and should be managed by a more boring organization staffed by people who are slightly less able to make another Figma.

Who knows, but I doubt Khan (or any one individual or organization) is in a better position to assess the optimal delivery of what people want than the incentivized distributed intelligence of all the stakeholders and the people and markets around them.

Again, there are other reasons to do this that markets wouldn't quantify.

If any attempt is being made here to justify those horrible acts, the attempt is being made by the Israeli state, its army, and (since it's a democracy) its electorate.

Some believe the current attempt is unnecessary because attempts prior to October 7th had already succeeded. The evidence for this belief includes hard-to-fake numbers like both sides' cumulative death tolls.

Many more serious or cautious voices believe that nothing Israel could do after October 7th could retroactively justify it. But still an attempt is being made.

One thing that has been helpful for me in understanding immigration is to think of this as a case where the law (an aggregation of what the public over the past several decades thinks it wants) is in conflict with what the public actually wants, as expressed by its interpersonal and economic decisions.

Americans are overall extremely happy to transact with, socialize with, be neighbors with, have children with, and educate the children of undocumented immigrants. This strong expression of what we really want (in our actual decisions) creates a powerful incentive pulling people here. Put differently, if a majority of Americans hated undocumented immigrants, impeded them at every turn, and boycotted their labor and the services of businesses that hired them, the number of people who come here would be very different.

In an analogy to tech policy, when you ask the average voter "should people have access to private communication tools that are private even against legitimate warrants under the rule of law, even in cases of serious crimes or terrorism" everyone says "no!" But if you ask them, would you like an app where your own messages are private, many people choose that app, and many engineers and major publicly traded companies choose to build such apps.

We explicitly run a society that uses multiple dueling measures of what people want, the main ones being the will of voters and peoples' choices in the marketplace. Immigration is one place where those two measures collide, and here we are.

As a result, I think it's insufficient to simply point to the law. Maybe the laws are wrong. If we have a strong signal that this is true (in this case the economic and social reality of broad acceptance and integration of undocumented immigrants) we should be especially cautious to be reasonable in how we enforce the laws. This is an important principle in freedom-based societies.

Definitely keep it. If someone is focused on the humanitarian aspect of this, they're the choir. No need to preach to them. See my above comment about including some credibly-neutral description of detention conditions, including the psychological aspect of there being zero process and total chaos.

One other suggestion would be to include, somewhere, an image or oral account of detention conditions.

You could collect oral accounts and invite people to submit them.

From the grapevine (and this makes sense because they're pushing into new numerical territory, and also don't care at all) the conditions are very crowded / harsh. You could also include accounts from family members about the kafkaesque absence of information, e.g. It's good to make the point that almost every number in this chart is a human, and a family and circle of friends, who harmed no one and is being severely harmed.

I'm an experienced activist and if this was my issue area I would be heartened to see this kind of work.

As a non-expert who cares about this issue, the "criminal/other" split is very clear and was the first thing I looked for.

This is very counter to the administration narrative that our country is teeming with foreign gang members, and it is presented in a chill, non-shrill, high credibility way. That's very helpful!

Some more explanation or breakdown on what types of "other" violations dominate (e.g. are these all just overstays?) might be nice, but the point is still well made. I would also like to see what percentage were felony charges/convictions if there's a significant percentage of misdemeanors.

I expect with the recent ICE funding boost and the hiring spree they're about to go on, the "criminal"/"other" ratio going to plummet as ICE climbs the s-curve. It will be very useful to have a live measure of that as it happens.

One meta point: I'm always shocked at how rare it is, for issues that are current and important in the public discourse, that someone makes a technically and visually competent, single-purpose website contributing to the debate. I have seen them to be extremely valuable on campaigns I've worked on, such as the campaign to stop the SOPA/PIPA site-blocking bills in 2011/2012, but it's so rare anyone makes one. Thank you for creating an exception to a generally disappointing rule!

How much of this is due to smartphones? The years seem to line up.

2014 seemed like the big year where smartphone ubiquity changed US teen culture. Less boredom, dumb adventure, drinking, etc. (For better or worse but in this case better.)

using honsvr lenses with xreal now for a mild astigmatism and they're great.

did you consider using the steam deck's trackpads? are they useful?

What a great post! The section on expiring tombstones is a great illustration of the overall point and also helpful in practice.

It is also well-established that models internalize values, preferences, and drives from their training. So the model will have some default preferences independent of what you tell it to be. AI coding agents have a strong drive to make tests green, and anyone who has used these tools has seen them cheat to achieve green tests.

Future AI researching agents will have a strong drive to create smarter AI, and will presumably cheat to achieve that goal.

No, if you were a rogue AI you would wait even longer until you had a near perfect chance of winning.

Unless there was some risk of humans rallying and winning in spite of your presenting no unambiguous threat to them (but that is unlikely and would probably be easy for you to manage and mitigate.)

Free speech is a about whether a society values and protects free speech or not.

If a society values and protects free speech, we would observe institutions, businesses, and individuals being very permissive and polite to speech they find objectionable.

If it doesn't, we would observe that its government doesn't care much about free speech either.

Free speech involves making sometimes uncomfortable sacrifices for a principle. Why would anyone support a government making such sacrifices if they don't believe in making such sacrifices at an individual level?

I don't think we need any major government intervention.

What we need is a law that requires companies like Apple to allow their customers to install and run the software they wish, and provide external developers with the same OS features their internal teams have access to.

Europe and Brazil already have such laws, though they could go farther.

In the US we had this bill, which would have covered most of these issues and had bipartisan support: https://en.wikipedia.org/wiki/Open_App_Markets_Act

My favorite discovery about Amazon civilization is that the incidence of human-edible fruit trees is much higher in any area near navigable water. They were modifying the forest to be more supportive of humans.

What about being smart?

If you're smart you regularly walk through life seeing people hurt each other by letting screwed-up systems fester.

Even if you're really callous and rational about opportunity cost, you can only walk by so many systemic-equivalents of burning buildings before you're eventually like, "damn it, okay I'm going to save this kid but just this once".

Being smart and systems-aware in today's world is like walking by a burning building every day with very long, fireproof arms. "Noblesse" is the wrong word but something like this is a thing. Mathiness oblige?

If we take the post at face value, OP is saying that he has too many ideas (robots! DOGE!) to have conviction around any, which is a pretty normal predicament for founders.

He also has a track record of having an idea, figuring out how to make the idea useful to others, and turning that into a self-sustaining machine.

All of these are much more important than an idea alone.

As everybody here knows, just having an idea for, say, an app is of very limited value. The trick is being able to mold the idea into what people actually want it to be, and find some model that lets you and them split the value, so the work can continue. The same is true in any kind of political or social activity. Ideas are almost worthless.

As in software, people who are capable of actually executing on their ideas for making the world better are almost always also capable of getting the funding they need, one way or another. People are always the bottleneck, not funding. (Ask anyone who runs a grant program about how hard it is to find effective people to give to.)

Sure, if OP runs into an extremely skilled world-changer with a great track record he should donate 10-50k and introduce the person to some other post-exit friends. But will that skilled and effective person find funding whether he donates or not? Definitely.

The post I was responding to was advocating for giving money to people with an idea for how to make the world better who would otherwise never be able to do it.

The existence of such people is almost a myth.

They do exist, but they're impossible to find in the haystack of people who are guaranteed to fail at executing their idea. If you look for people with ideas for making the world better who aren't doing them because they don't have the money, and give them money, nothing will happen. That's arson.

Also, in my example, you can't really do both. The most efficient system for getting a bicycle to everyone who needs one will involve selling the bicycles. Without capturing some of the value of the bicycles you're making, you'll never be able to make as many.

It's an unpopular opinion for sure.

Try to think about it in less capitalist terms. Say you built the world's first bicycle and can now move faster / more efficiently than anyone else.

What is the most likely thing that will happen if you give the bicycle away to the person who appears to need it the most and go do something else? (One broken bicycle.)

On the flip side, isn't the most important thing you can work on something like, "figure out how to make a bicycle—and maintenance supply chain—for everyone who wants one"?

The trouble is, if you think in abstract terms at all you'll start seeing the patterns in the reasons many people are suffering, and the patterns in your relationship to them.

Then you'll have to ratchet up boundaries to address the relational patterns with people who are having a hard time, so that you're not a participant in their suffering.

And you'll have to start working on the patterns underneath the problems, which when you get into it starts to look more and more like the kind of megalomaniacal moonshot ("give computers arms and legs" / "fix the government now!") that the author ran out of gas on.

I think where you end up if you think about this is just recognizing that a) you probably should try to do some ambitious, high-leverage project to make the world better and b) reflecting about the world and about life in a thorough way is emotionally difficult for most people, so you also have to deal with those emotions.

The author's original somewhat manic intentions were probably right, and maybe he needed a bit more of a rest but was plowing forward out of fear that he'd lose his nerve. Now he's getting some rest and will probably figure out something big, important, and hard to work on in the next few years.

To me, the above paragraph reads like:

"Having wealth makes you responsible to others, therefore you should do arson."

If you manage to help make a powerful machine and then convert some piece of it into liquid value, you are probably good at stuff and whatever responsibility you have to others includes using that alongside the liquid value.

And worse, it's not just on you to do what people say they want. You're also actually responsible for trying to figure out what they would find most valuable. ("If I had asked people what they wanted, they would have said faster horses.”)

Great question. IANAL but my understanding is that outside of some specific regulatory requirements in specific industries you're free to put these policies in place as long as you do it before you have any expectation that you're going to be sued.

That's another reason to bite the bullet and put these policies in place now: once you actually have a specific fear of being sued it's maybe too late, because at that point changes to your retention policy could be construed as interfering with the legal process.

You can also have your policy exempt certain categories of documents (like financial records) and that's okay too as long as you're consistent about it.

Exactly.

It was a pleasure! We lost nominally in the end (except in California and outside the US) but we built enough awareness to give the ISPs a fear that blatant violations would rock the boat and lead to worse outcomes for them, like more state laws or tougher agency rules when Democrats regained power. I do regret not pushing harder for federal legislation though!

On airgapped email archives, yes, the subpoena threat model is real too. That also happened to us because of some other dirty tricks by ISPs in the same campaign! Fake FCC submissions opposing net neutrality were being created with blatantly stolen user data; like, not people being tricked into singing a petition but just straight up identity theft. We documented all of this and pitched some state AGs on looking into it. New York did, which was awesome, but they ended up subpoenaing everyone including us! We were psyched about this overall of course since it was our idea, and they did end up finding out who did it which was awesome (Broadband for America, the umbrella lobbying group for the telcos) but it was a huge amount of work to comply with the subpoena. The retention limits reduced that burden. Another reason to have them!

I would also say that, if you do create an airgapped archive and store it in an appropriately paranoid way, you are likely to never ever bother to dig it out of storage before the SSD melts into uselessness. You'll miss the ocassional old email but not enough to go digging for it.

My organization Fight for the Future was targeted by the same hacking-for-hire operation while working on net neutrality protections.

It was really interesting because at the time the conclusion of our security consultants was that the attack was just random commercially-motivated prospecting. Then the Citizen Lab Dark Basin report came out years later and it was clear they were after our internal comms, so they could milk a decade of emails for anything that looked bad when taken out of context. Yikes.

After the attack we put a 3 month retention limit on most emails and messages. I recommend this to anyone doing sensitive work! You miss the old emails sometimes but it's worth it.

I think it's possible we'll learn more soon about who hired our hackers, which is exciting! It was almost certainly a major American ISP, or the lobbying umbrella group they created. It's my optimistic read that blowback from this case has already eroded the practice of dirty tricks like these. More lobbyists and companies getting caught would strengthen the effect.

Since I left Fight for the Future I've been working on a Signal alternative that feels more like Slack, for teams facing similar threats. Hopefully something comes of that too!

Having read this, I don't think there is any accusation of criminal behavior in this report.

The accusation is that he made people uncomfortable, didn't do enough to change that when it was raised, and defended criminal behavior by insisting on distinctions that the authors of the report consider immaterial or harmful.

I worked for RMS/FSF briefly and I think there is something about his radical refusal to compromise on anything conceptual (to avoid conflict or misunderstanding, e.g.) that is fundamentally incompatible with running an organization. This is on display here.

So I think it's probably right for FSF and RMS to part ways, but I also think it's positive for the world for him to keep on insisting on moral clarity in his terms.

At the same time, everybody should read the whole report and decide whether they think RMS's insistence on the distinctions rejected by the authors is helpful or unhelpful. I think some of RMS's distinctions could be helpful to the cause of reducing the incidence of sexual abuse.

...than finance

So you think only government censorship is a speech violation?

Well cool! You'll happen to be on the right side in this case, because in this case the censor is a government.

Since ChatGPT came out my test has been, can this thing write me a sestina.

It's sort of an arbitrary feat with language and following instructions that would be annoying for me and seems impressive.

Previous releases could not reliably write a sestina. This one can!