Hopelessly optimistic. Spoken like someone who has never taken an experimental design course imho.
1) The search space is practically infinite dimensional. All methods suck when trying to extract a causative model from an infinite search space. There's a reason why in experimental design we change as little as possible.
2) SNP's are not the entire story. If it were this simple we would have progressed much further already. See dismal failure of all other high throughput sequencing and microarray technology. We still don't know how to analyse this stuff properly, if it will ever be possible.
3) The metabolome is adaptive! While we each have different enzyme kinetics due to slight differences in protein makeup, overall metabolic flux rates are amazingly consistent. See Oliver Feihn et al for more details.
Sold instructions on ebay about how to beat a little known tax law [it's now plugged]. A friend went to work for the tax man and in training they were told about me and how it's so important to not let these tricks get out, as it could cost the country lots of money. I promptly stopped as I didn't realize I was on their radar!
Reverse engineered an algorithm from a huge international company who kept it under lock and key using publicly available data and used it with their competitors. They still have it under lock and key.
Bought textbooks at fundraising sales and sold them to university students the next year for a 3000% markup.
Let's say you had an eclipse plugin that let you write in lisp, but it actually converted the code to java underneath. You could click a button to switch between views. Are you saying that would be a bad thing? If a tool allows you to be more productive, I use it. Note: I don't use lambda4jdt at all, I just like the idea of creating problem specific views for code and I don't think that belongs at the language level.
The smartest people I know have the following in common:
- They are the most knowledgeable person in the room in their topic of interest. That takes obsession.
- They surround themselves by other smart people constantly
- In large groups of smart people they are perfectly happy to ask questions even if they are wrong. E.g in a maths seminar will debate, and often get beaten by the subject expert.
I often wonder if the latter - asking "stupid" questions to experts in a field - is something that was there before they were at the top. Anyone else noticed this?
Even though code is text, you would think someone would have tried to build a different view given the popularity of MVC.
Inline expansion would be nice. So would code paths. So would inline images & html rendering for documentation. So would the relevant rendered html for the issue relating to the code you are looking at. So would other annotations. And what about visually hiding code when working in a language like java?
I tried to get this going in eclipse once, and it really wasn't going to happen easily so I gave up. I've always hoped someone would build something similar as I feel it would really help productivity having everything relevant right there with the code.
Note: The project http://code.google.com/p/lambda4jdt/ is the closest thing I have found to what I am suggesting which I think shows the power. It just goes to show that it could work.
I never said anything to the contrary. I said that it's a false dichotomy; just because you don't spend time doing something doesn't mean you don't like it, just that you may not like doing it as much as other activities.
I love playing games, but never do because I like other things (learning mainly) far better. I also love kayaking, but would much prefer to go to the gym because of convenience. That doesn't mean I don't like playing games or kayaking; I really enjoy doing both. Get it?
Honestly, the world isn't black and white like that. And I can see why you argued about it; that would drive anyone mad being told what they do and don't like based on a histogram of time spent in the last 2 years.
There are many explanations for not spending time doing things you like; competing priorities is the main one (i.e. I like other things more than them) and also environment (perhaps you enjoy kayaking but hate the drive to get there, or don't want to go with johnny and max because you don't like max).
Just because you don't do something doesn't mean you don't like it. It's a false dichotomy.
Once an itch is scratched it turns into a scab. And new itches appear.
Seriously though, businesses change over time. Are people who are great at getting things going the right people to run them long term? In some cases yes, in others no. In my case I plan to be involved in one of my companies for a long time, but for another I am involved in we are building to exit.
Years ago (pre swivel, post quantrix/tableau) I made a competitor as well. We quickly decided once we went out and talked to potential customers that it was a no go.
1) Very few customers wanted to upload their data to the web. Even less were allowed by law.
2) Most had integration problems - i.e. they couldn't actually access their data because it was in disparate systems.
3) Most people didn't actually use the information for anything, it was simply to provide them with "evidence" that they were right before looking at the charts. As a statistician this makes me sad, but it was the way things were at least when we investigated it.
Over time we gradually realised why BI was priced so high; it needs system integration alongside it. Now it turns out that we may have been wrong (gooddata.com are doing a good job as far as I can tell) but I still think it's the tip of the iceberg in terms of potential market.
My advice: target a niche with a huge problem where the customer is unable to get at their data at the moment. Choose the niche which has 1-2 big vendors of data collection systems with 80+% combined market share and provide the solution to the problem. And charge.
You need enough knowledge to know when someone is making the right decision or not for your business. I run a successful software company, have hired and managed many programmers and still got it wrong with hardware.
>I assume the next time you see an opportunity in the hardware business and want to jump onto it, you wouldn't be taking a 4 year undergrad course to know the drill.
That's actually exactly what I'm doing - although after you have a degree it only takes a year or so to go through the relevant programs, and if you know people in the university it's easy to get private tutoring for specific gaps. It's not like the courses are difficult once you learn how to learn. Plus, you get to see who would be a good fit in your next startup.
Learning by doing: Perhaps I'm just not smart enough, but I need to understand some theory before I can do it right in practice, and I learn best talking to knowledgeable people. I fumble and make too many costly mistakes when learning something completely new without guidance. We all have our weak points :)
The odds are stacked against you if you don't know the area you are in. This is because you can't control the competency of the people you hire or the work they do, so you have a random factor that has a direct and long lasting influence on your companies success.
After my first software startup I tried a hardware product and failed miserably because of my lack of knowledge. To remedy this for next time I am formally learning the areas where we failed (electronics + cad) and when I'm ready I will try again.
Theres so much that can go wrong with a startup that it's just not worth adding another element into the mix.
Take the intersection of the three things you know most about and are passionate about. If you're not sure about this simply look at your most visited sites.
Then use those are your "base" and try to find ideas that combine all three.
Well, people claim that digg is entirely run by the top users from what I understand, and if only 6/10 stories that make it to the front page are predicted by the top users voting patterns, theres not much to that claim right?
6 out of ten stories in the new links section can be predicted by whether the top users vote on them. Really, only 6 out of 10? This seems extremely low given the claims of gaming (and if I'm understanding it correctly).
Someone should create a "everything" test, which pulls graduate level information from every topic maths, statistics, physics, chemistry, biochemistry, physiology, medicine, therapy, psychology, english, history, art, accounting, business, marketing etc etc.
I'd love to take it just to see quantitatively how much I don't know.
My goal when I first started was to exit and buy a fully equipped bio/physiology/elec/engineering lab and enough raw material to invent full time.
I thought I'd need 10m+ to do this, but over time I'm gradually acquiring everything I need due to places like sparkfun, ebay and alibaba for little more than the cost of an average car.
I'm inventing every day (5-6days software, weekends building random projects), learning more than I ever thought possible and loving every minute.
My point is that if this is anything like your dream, it's readily doable now as the cost of doing science is extremely cheap these days. And I've found that productivity increases when I take a break from software in the weekends and do some wet/hard projects.
1) The search space is practically infinite dimensional. All methods suck when trying to extract a causative model from an infinite search space. There's a reason why in experimental design we change as little as possible.
2) SNP's are not the entire story. If it were this simple we would have progressed much further already. See dismal failure of all other high throughput sequencing and microarray technology. We still don't know how to analyse this stuff properly, if it will ever be possible.
3) The metabolome is adaptive! While we each have different enzyme kinetics due to slight differences in protein makeup, overall metabolic flux rates are amazingly consistent. See Oliver Feihn et al for more details.