I think grandparent comment's point was that at some point, greater good for disabled people can be delivered by specifically helping them rather than re-engineering the world to fit them. Considering that each has a concrete price tag attached to it.
The "wheelchair that can climb stairs" vs "adding ramps everywhere" comparison.
I feel like in the past the latter made more economic sense, but at some point technology might tip the balance the other way.
Why are you primarily looking for >billion dollars startups?
I would wager that if you take a look at an accurate distribution of markets according to potential you would find a magnitude more small/average markets (that are untapped) than billion dollars ones. And similarly, I suspect that the success rate for those "boring" ventures is much higher than the exciting shiny rising stars.
Question: Why not optimize for companies that are certainly not going to become Airbnbs but will capture the full value of an averagely sized market (say between 50 and 300 millions)? And if my guess is correct and they end-up eating a lot less resources than the soon-to-be-unicorns, you could even optimise for volume.
Is the pay-off (wrt. to the energy spent and success rate) for unicorns really worth it?
1) This is where effectively all of our returns come from.
2) These companies are the most fun to work with.
3) It turns out that it's harder than it sounds to capture the full value of a smaller market for a bunch of reasons, and so the failure rate is much higher than expected.
I think the payoff, at least in our case, is well worth it.
>3) It turns out that it's harder than it sounds to capture the full value of a smaller market for a bunch of reasons, and so the failure rate is much higher than expected.
What kind of reasons? Direct me to some resources if there are too many to enumerate
Primarily because it's hard to pivot in a small market.
When you are in a massive market there are lots of adjacent places to pivot. When you're in a small market, if you miss on the first try, you probably miss completely.
Things go wrong in business regardless of market size -team issues, tech issues, funding issues, operational issues etc.
Your probability of success in a small market is going to be marginally better than it is in a bigger market. So if you look at it in terms of expected value, the bigger market is more valuable given the probability of failure.
I'm interested in this as well. Doesn't this go against Rob Walling's advice in "Start Small Stay Small" as well as Peter Thiel's advice in "Zero to One"?
I read Peter Thiel's advice as being more along the lines of finding a small beachhead that you can conquer to get started, then using that as a base to push into a big market.
The current situation in France is making me very conflicted about privacy and public liberties.
I am a strong believer of the value of privacy for a modern society. I even want to dedicate my life to help building privacy enhancing technologies and censorship-resistant networks. Because I think that an "advanced" world can only thrive if information is unrestricted, or unstoppable.
But today I face a dilemma. The dilemma of choosing between freedom and privacy, and security.
I am French, and have a lot of family in Paris. My brother lives two streets from the Bataclan and lost one of his friends. Some of my friends lost 5, sometimes 10 people that night. Imagine loosing two thirds of your group of friends in a few hours. This is frightening.
When I look at France. I see a great country, with a lot of humanity and when I look at the French, I see a freedom loving people who share a love for good food, good music and generally speaking, the good things that life has to offer.
But I also see the failure of my country in the suburbs. With entire neighborhoods that have been left uncontrolled by the government at some point, and who never went back to that state despite lots of efforts. These neighborhoods are rigged with crime and violence, and have been a fertile environment for religious lunatics to grow stronger for the last twenty years.
And I have mixed feelings. The French National Assembly has extended the state of urgency to three months. Strengthening the regalian power of the state and weakening the counter-balancing power of the Judicial branch.
Hundreds of raids have been coordinated through France, most being in those "uncontrolled areas". And it seems to work.
Which prompt me to think that this might be for the better. For the most part of my "short" life, I have thought that a people should never "trade freedom for security". But I have come to the, perhaps wrong, conclusion that there can't be "freedom without security" either.
Maybe we should give up some freedom to let the "good guys" crackdown hard on the "bad guys".
But maybe it isn't. Maybe fear is clouding my judgement.
I'm very sorry to hear about the losses your friends and family have endured. My comments below are not meant to dismiss those losses.
So, here are some questions to ask if you want to prevent fear from clouding your judgment of the security vs. freedom question:
If the authorities knew where they needed to raid already, why didn't they raid there before the attacks?
On the other hand, if they have all of the surveillance we read about in the news and more, why weren't the attacks prevented? How will more of the same surveillance work any better than the massive amount already in place?
Finally, how do you trust the authorities claims now that their emergency techniques work, if you didn't trust their techniques before?
Again, none of this is meant to belittle the losses sustained or the seriousness of the attacks. But I believe that terrorism is a social and political problem first, a criminal problem second, and a surveillance problem least. So the solutions need to be long-term social and political moves to counteract the economic and ideological conditions that breed terrorists.
I do understand your viewpoint – theres an unfairness in large-scale terrorist attacks that's very hard to deal with, and commentators can all too often be a bit shallow in dismissing that.
The thing is… there's little evidence that limiting privacy would prevent terrorist attacks. Maybe unless you are talking about a complete elimination of all private communication—something which I think most people would rightly reject outright—then there will always be private channels available for communication. I think most people will accept that some form of government access to communication is sometimes warranted – but that such access must be strictly controlled, judicially overseen, and limited in scope and time. Previous experience as a society suggests that any excessive government power will be abused.
The issue here is not one of a government which is limited in its power to 'crack down' on terrorism. It's what you've identified – neighbourhoods riddled with poverty and crime. There are a breeding ground for disenfranchisement, and we know it's the tactical approach of groups like Daesh to encourage and recruit from these disenfranchised areas.
You can only realistically solve this problem by solving disenfranchisement. That's a complex socioeconomic issue, though – and not something I have the answer to. But the idea that this is 'freedom vs. privacy' doesn't stand up to examination.
I'm really curios, why do you think that less privacy and more government control means safety?
Suppose that this less privacy equals greater safety, is it smart allowing the government to have this sort of power over us? Sure maybe the current government is awesome and really wants to keep us "safe", what if the next president is an evil tyrant who suddenly has all that power.
People can ultimately use any tool to do evil imho, would you like the government to control who can buy knives and cars for example? both of them can be used to do evil acts.
I wrote this to Pavel Durov, and maybe it will help bring some more clarity to the freedom vs security debate:
Listen, as a fellow social platform developer and someone who has to seriously consider these issues, I have a suggestion. Why don't you consider terrorism as an epidemiological model, using your network to spread a certain undesirable virus/meme (whose symptoms include violence). Then you can borrow from known bio literature about battling viruses:
Markers (based on previous samples
Antibodies (preventing recipients from getting infectious messages)
Obstacles (making difficulty of spreading organically harder)
And work to promote herd immunity. You can add this to the system without any government or yourself being aware of the contents of the messages. Have the computer system be like an organization that maintains herd immunity in its population, while at the same time (being electronic) having zero interest in reporting people and message content to others.
One of the first things you'll notice is mass media (one to many transmission taken to the extreme) is the biggest risk to herd immunity. Free speech like every other right may have limits... speech to 100 people may be totally unenmbered but speech to 1,000,000 comes with responsibility, or else it's just a freerider on a system. YOU Pavel have a responsibility since your code is your (meta-)"speech" once such a large audience engages with your product!
Stop shutting down ISIS channels. It's GOOD that they use your software. Look at the incident of Craigslist shutting down its prostitution section -- people just moved elsewhere. Instead, think about whether "right to free speech" and "right to privacy" must be absolute or whether there is justification to BALANCE them against public welfare at ANY point on the continuum. Perhaps if the system can detect with good confidence some violent speech, it is better to LET IT HAPPEN and alert the operators so they can refer it to the authorities. Either way, Pavel, as long as the developers are few they have the power and can be bullied by governments. So use this power to set the lead: build VIRUS DETECTION not BANNING into the system, and only drop privacy in the case of some imminent danger. Make the networks have an immune system. But do not simply make people move somewhere else.
You seem to believe that the EU is one homogeneous, federal entity. It isn't, yet.
Having the EU parliament vote a motion in favour of welcoming Snowden as a political refugee is great. As the article highlight it isn't binding to the EU commission or EU countries. But this could help document Snowden's in front of the ECHR.
The European Court for Human Rights does have supranational power over EU states.
This is a small step in the good direction, no need to be cynical.
> You seem to believe that the EU is one homogeneous, federal entity. It isn't, yet.
I did say 'EU country' didn't I? So why do you think I believe that the EU is one homogeneous federal entity? I'm pretty well versed in EU politics and I'm kind of surprised that you could read that into that line.
> Having the EU parliament vote a motion in favour of welcoming Snowden as a political refugee is great.
Yes, but it amounts to exactly nothing without any particular country taking actual action, the EU parliament has a long history of being ignored on issues like these in spite of well meaning words their ability to exert pressure on the nation states to actually act is in this particular case quite limited. But it's a nice gesture, I'm sure Edward Snowden will appreciate it.
Talk is not cheap at all, my dear friend.
We europeans are going against US requestas, even though We are technically allies.
Do you think dieselgate has been discovered only now in US by accident?
I think that can be true. But I think that an excellent virtual course is still inferior to an excellent physical one.
I mean, 90% of the interesting stuff I am learning during my classes comes from the discussions that start at the end of the course with the professor and a few others.
I do enjoy the ability to rewind, and choose my pace when taking an online course. Too bad the quality of the interactions I have mentioned has not yet been captured by MOOCs or OCs.
Agreed, but I'd counter and say that a median virtual course tend to be better than a median physical one. The more advanced the course, the more this tend to be true.
Lecturers in a classroom, specially professors who's main goal is to research and are forced to teach, just come and blurt out whatever they think today session is about. I once had a class where the slides had scrambled egg over because the professor had been preparing the class over breakfast, with his baby boy on his lap!!! And I am not talking about an overworked postdoc, this was a guy with tenure. I imagine he though he was being brilliant by cramming together his parenting, contractual responsibilities to students and fulfillment of bodily needs into a single time slot so he would have more time for his oh-so-precious research.
When this kind of lecturer do some MOOC course, they become conscious there's going to be an audience and a record of their performance. So they devote at least as much effort to it as if they were writing a paper for some mid-tier congress. This has the double advantage of having them put an honest effort, and the delivery medium playing to their strengths. This shows in the end results.
I am working on an implementation of Raft and that's a gold mine of issues to look for. It's also relieving to see testing tools for DSs coming up.
DEMi is written very clearly, and it's a pleasure to read the code: https://github.com/NetSys/demi . Definitely challenges the research "spaghetti code" stereotype.
I was in college in 1998 when the Starr report came out, detailing President Clinton's sex life. Back then it wasn't easy to download such a big document to your computer, so a lot of people came to the computer center, which I managed, to look at it. But it was long and they didn't want to read it in the computer center, so they started printing out the 90+ page report! (printing was free)
It go so bad we had to ask all the people that printed it if they could bring their copies back when they were done, so we could have a lending library of the Starr report.
My point is, you're right, it's a great time to be alive -- you don't have to tell anyone about your interest in these things. :) (although on the flip side there was a pretty good watercooler discussion of the report at the computer center)
I don't think anyone should be surprised that an intelligence agency - that has repeatedly violated its own country's law, and actively contributed to the weakening of civil rights - be guilty of this sort of negligence. That is exactly what happens when an institutions is allowed to grow unchecked, with no or little civilian oversight or consequences for the wrong-doings.
What's scary is that this kind of clueless, and technology illiterate, people are actively involved in shaping the future landscape of massive data collection.
I think we are about to witness, in the next decade, multiple "incidents" where millions, perhaps billions, of private records about innocent citizens will be leaked because of this kind of negligence.
I think people do deserve to be surprised. Competence is not the same as selflessness. Many people routinely question whether the FBI is operating for the good of the country, but most people at least believe that they are good at their job.
I'm not sure that the CIA has ever been a real ally of the people in general, but to the extent that they are, it's similar to the inclusion of the USSR in the Allied Powers during World War II.
Nice rant. The last two paragraphs threw me off, though... how exactly is the CIA Director involved in shaping the future landscape of massive data collection?
In addition to that, I wager there's value in extending the current infrastructure so as to not ostracize disabled people even further.