I've notice some non-native English speakers use "double verify" where native speakers would use "double check" or "verify". It could be a play on that.
Or they had a meeting where the founders asked "what domain sounds twice as good as verify.com?"
If you are a market maker / arbitrageur working across multiple exchanges, US dollars can take days to settle and wire/withdrawal capacity is limited. USDT offers instant transfers, so you can rebalance across locations with ease. That's why it's being heavily used by the largest market makers in crypto.
I think you've answered your own question. Taking outside money will make life busier, increase pressure, and create investor expectations that may be different from your own.
Bringing in an investor that would be willing to replace you with a new management team is a Bad Idea(tm). It's also a bad idea to try to mislead your investors and replace yourselves.
Outside money can be great if you want to build a really big business and need the capital to win the market. But there are tons of businesses that are better off using revenue as funding.
One of the drawbacks of the tech media is how few people talk publicly about the businesses where the founders would have been much better off if they'd only avoided funding.
Fog Creek is an interesting example partly because it's written up and also because it's kind of a hybrid where they've remained self funded but spun off VC funded ventures in Trello and in some ways Stack Exchange http://www.foundersatwork.com/joel-spolksy.html
I work full time in a Head Start agency and work with over 250 students every week in a STEM Lab for over 8 years. I am proud of what I do and even more proud of the teachers, students and families.
1) "Teaching to the Test" is to teachers. They are tested on CLASS (Teacher engagement with the students) ECARDS (Classroom environment) NAECY (Classroom and Teacher Lesson Plans). Federal Reviewers check for Food and general Safety.
Children are engage at their point of interest. So if many in the classroom are interested in a subject the teacher will structure the classroom to focus on that subject and make projects etc..
2) Parents are very much apart of Head Start. Every one has a Policy Council that parents actually help run and make classroom policies. We also have Family engagement partners that visit the homes as close to monthly as possible to help the families to create successful students. The teachers are required to visit the homes at least 3 times a year.
3) I think that is the issue. K-3 does not go through NAEYC or CLASS and usually developmentally inappropriate. There is no family engagement in my experience.
CURRENT POLICY of Head Start Agencies - The lowest 10% of CLASS scores are to be de-funded and the contract to be put up for competition. Also any missing children or serious safety violations equals immediate de-funding.
"teaching to the test" depends on the instrument they are using - I cannot find it in the study.
We used LAP and ELAP (8 domain assessments) because we started much earlier (6 weeks). It is structured so you do teach the test but when the test says "stack 3 blocks on top of each other" (obvious early LAP skill), you are pretty much going to teach the test.
In my experience (90s research program), this is basically in line with what Harvard found about Head Start. Starting earlier has the effect of finding learning defects (not sure the current term) earlier like hearing problems and allows time for correction during the critical 0-3 years.
2) can be discounted probably, because students were chosen at random
1) and 3) seem a stretch. How do you teach to test in such a way ( assuming no cheating by teachers ), that kids can do math on the test, but later they lose that skill in such a way that they become worse (!) in a statistically significant way than those who weren't in pre-k ?
I think what it is, is bad sampling procedures. For example, in the study, they had to get permission of parents to evaluate the kids and only 36% ( ! ) of those not in pre-k agreed to be evaluated. So, I think this is just biased sampling at work.
Having looked seriously into this study it does appear to be done correctly. I struggle with it is Pre-K's fault and not what are we doing wrong with K-3, which is were I think our Public Education System (Especially in Urban Schools) are at our weakest. The "common sense" of start early for reading became all 5 year olds must learn to read became my daughters 3 hours of reading a day Kindergarten and 2.5 hours of math with no recesses.
The issue is not that it's done incorrectly, the issue is that the limitations of the available sample are such that there will be a bias introduced. Most parents did not respond to request to be evaluated ! that's a very strong indication of a possible bias
It states that the raw values come from the school database with a sample size of around 3,000 children. Also the teachers were giving the responses at the end of the school year.
The parents did have to sign a letter at some point to be apart of this but it appears that there is an issue with the outcomes. I fully believe what is happening in my Head Start really is making a difference, but it seems to be harder to see the advantages right now with this research paper.
I'm not a lol player, but I'm amazed to see a similar bug that existed on early versions of Battle.net 15 years ago. However this didn't had many impact since we played top vs bottom with game assigned positions (starcraft).
We'll never know for sure. There have been many times where I'll type in chat: "top", my message will appear, and then somebody else will say "top" immediately after. But, a asking a third person reveals that the second player's message actually arrived first, even though my client shows my message first.
Generally speaking users expect instantaneous feedback. It would feel super shitty if every time you hit enter there was lag until a server roundtrip posted the message.
Messages that the server saw earlier could easily be inserted partway up, right? We're talking about ~100ms most of the time; that's not going to cause you to miss messages.
Some chat service I use does this and it's confusing because I don't expect it. I check the last message to see if there's anything new, not necessarily the scrollback.
Basically, it triggers an annoying O(N) operation whenever it happens :)
> It would feel super shitty if every time you hit enter...
I and many people I know don't agree with this. Instant feedback is nice, but not required. Moreover, if we need to choose between (instant feedback) and (knowledge of server receipt of a message and knowledge of message ordering), we'll choose the latter every time. (I'm glaring at you, Google Hangouts.)
Having the chat server lie to you about the state of the world always feels far worse than a delay between the time you submit your message and the time it appears.
Maybe the best way to claim top lane shouldn't be the first person to type "top" and hit enter.
Even if the server corrected the order messages were received it'd feel shitty to not get your lane because someone has a faster roundtrip to the chat server.
1) The likely worst-case trip time to the chat server is 250ms. [0] It's true that that's ~2.5x slower than the typical minimum human reaction time, but it takes a fair bit longer than this to type "top<enter>". [1]
2) The case where the server displays one message ordering to one player and another message ordering to another player is strictly worse than the case where the server takes 250ms longer to display the same message ordering to all players. The first case creates confusion and arguments, as well as feelings of distrust directed towards the chat software. The second case might cause a vague sense of displeasure at the quarter-second delay.
> Maybe the best way to claim top lane shouldn't be the first person to type "top" and hit enter.
Agreed. Perhaps people could be civil when they play their games. :)
[0] In this case, RTT doesn't matter. Only the time required to get a packet from you to the server matters. I think that we can reasonably assume that the worst case RTT from anyone who's playing a vaguely-latency-sensitive online video game to anywhere else on the Internet is 500ms.
[1] Honestly, it might take half of this time for the pixel issued by a draw command to your graphics card to appear on screen. [2]
Finally! I haven't played in over a year, but when I did, there was no correlation between the order of messages I saw and the order of messages somebody else saw... made for some rather frustrating arguments.
Although that matchmaking change where you get to say where you want to go and it matches you with a team helped too.
Do you know which patch specifically? Because I'm pretty sure I've seen out of order chat delivery in the past few weeks, although I suppose it could have been people being jerks in the pick phase.
One suggestion: On my first view, I ended up looking "backwards" for ~75% of the movie. I might have hit something at the beginning, but my forward-facing orientation couldn't see any of the action and there weren't any visual cues to tell me that if I turned around there were things happening until late in the movie. May be one of those weird things that VR needs to handle differently.
Excited to see more!
PS - I have a mac and DK2 but it's brand new, so could have been user error.
This actually strikes me as the #1 limiting factor for this technology and interactive fiction in general. A good story depends very much on causality. If the viewer in a virtual environment can get stuck in the corner (ie failing to direct attention/activity to key narrative elements) then the basic options are to a) move the viewer along automatically, as in some videogames or b) wait for the viewer to get un-stuck and pick up the story again. Unfortunately the first approach ends up more as a ride than an interactive experience, while the second baldy undermines the suspension of disbelief.
This is not to say there's no way to do it - if you establish the basic narrative grammar early you can certainly tell an interactive story, but probably at the cost of limiting complexity. The most effective examples I can think of are from ThatGameCompany, especially Journey.
Awesome breakdown. I wrote a blog post in late September - based on some industry rumors - that speculated on whether Candy Crush was "cheating" by varying the random seed to generate monetization or retention events:
Based on the "seed" going back and forth at the start/end of games, I'd have to assume that they are doing something with it. Anyway to see if that's happening?
They're probably generating the level with the seed. I don't think it's very useful, other than testing on their end (so they can check afterwards what the level was).
You could be right, but I think the seed is a lot more interesting. If you simply wanted to generate the level, you could use a random number generator on the device and wouldn't care about logging it.
As you showed by varying colors, the candy distribution determines how likely the player is to win, whether they fall just short of clearing the level, etc. This will have a significant impact on whether they pay to continue, buy boosts, etc.
They may not have done it at launch but as more and more people play and they collect the stats for each seed, I'm sure they can track how easy / difficult each seed is and send the seeds that they wish to do whatever objective they have.
Most of these F2P play games collect and analyze an insane amount of player activity and have sophisticated analytics that they then analyze to make the game more addictive or to increase player spending.
This is a nice analysis and good thinking, though it's worth noting that liquidation preference makes the VC equation less favorable. I'm not sure how Y Combinator works, but preference plays a shockingly large role when you run these types of calculations against your hypothetical VC deal.
Given PG's formulation, liquidation preference plays no part in the value-of-equity equation, since it can be factored out into the average outcome.
But sure, it does tend to depress average outcome. Then again, if your outcome is dominated by the presence of very-high-value possibilities, a reasonable liquidation preference may be no big deal.
