p********r
发帖数: 1980 | 1 我是做Business/Marketing的,虽然是技术背景,对machine learning/NLP有兴趣.
最近在创业团体里认识一个搞Deep Learning的博士(非中国人),说他想创业. 从技术上
他有两个想法,一个是Reinforcement Learning的应用不普及,他觉得这个方向有前途.
另一个是他在做的Deep Learning的研究,他说还没有发表,但是觉得很有创新,以后也可
以应用. 但是他不想去Google, M$当螺丝钉,想自己创业. 但是他不知道谁会买他写的
library.
我自己做的项目跟他有类似,但是少一个CTO, 所以一拍即合. 我说我来替他找Business
Model. 提了不少建议都给他否定了,因为他要做高技术,看不起普通的应用. 要有时间
研究,不能有Deadline. 不愿意跟不懂ML的人答扯,所以不能是用户接触高的项目.
最后想了我两个多星期,觉都没睡好,想到一个不错的方向, 几百亿的市场但是中小公司
都是Unserved/Underserved的. 我随机找了十个公司Email骚扰,给他们讲我的产品构思
,一天内两个有兴趣. 产品做出来一个是Subscription Model,但是维护比较低. 我又
有一些关系可以最初做渠道/客户. 结果我兴高采烈地给他看用户的热情,他说啊,这不
是说我的library要为这个市场customize,这个对我是太大的distraction了. 而且,你
找的用户居然是普通企业用户,不是Machine Learning Specialists啊. 我觉得还是卖
ML Specialists好,他们也都想用我的Library, 说不定也够钱支付我的开销了.
你这是想开公司呢,还是想找份科研工作啊? 我真是白忙了! 这个版上搞技术的多,你们
是不是觉得搞Business Model的不是忽悠就是拍脑袋随便能来一个的呀? 要考虑各方的
利益,渠道,客户是谁,长期竞争优势,自己有限的资源怎么分配,Revenue/Cost.我容易吗? | p********r
发帖数: 1980 | 2 One additional note. A colleague at my day job asked me yesterday, "You have
many good business ideas, but I can feel the kind of frustration you must
have felt that the management will not take most of your ideas".
Well, I couldn't tell him that I am trying start-up on the side. In the
start-up world, I will be lucky to only have five setbacks a day, I have two
for breakfasts each day. LOL.
At work, I at least got one idea executed (from market research, product
definition, partner selection, to system design and supervise testing were
my responsibility, I got help from others to contribute in many other areas)
that is bringing the company millions of profit each year. At startup, I
still have nothing to show for, but it doesn't mean any set backs will stop
me from trying.
.
Business
【在 p********r 的大作中提到】
: 我是做Business/Marketing的,虽然是技术背景,对machine learning/NLP有兴趣.
: 最近在创业团体里认识一个搞Deep Learning的博士(非中国人),说他想创业. 从技术上
: 他有两个想法,一个是Reinforcement Learning的应用不普及,他觉得这个方向有前途.
: 另一个是他在做的Deep Learning的研究,他说还没有发表,但是觉得很有创新,以后也可
: 以应用. 但是他不想去Google, M$当螺丝钉,想自己创业. 但是他不知道谁会买他写的
: library.
: 我自己做的项目跟他有类似,但是少一个CTO, 所以一拍即合. 我说我来替他找Business
: Model. 提了不少建议都给他否定了,因为他要做高技术,看不起普通的应用. 要有时间
: 研究,不能有Deadline. 不愿意跟不懂ML的人答扯,所以不能是用户接触高的项目.
: 最后想了我两个多星期,觉都没睡好,想到一个不错的方向, 几百亿的市场但是中小公司
| I****r
发帖数: 795 | | a*****1
发帖数: 372 | | l******n
发帖数: 9344 | 5 你找错人了
这种geek没有商业头脑，做不了战略伙伴。
.
Business
【在 p********r 的大作中提到】
: 我是做Business/Marketing的,虽然是技术背景,对machine learning/NLP有兴趣.
: 最近在创业团体里认识一个搞Deep Learning的博士(非中国人),说他想创业. 从技术上
: 他有两个想法,一个是Reinforcement Learning的应用不普及,他觉得这个方向有前途.
: 另一个是他在做的Deep Learning的研究,他说还没有发表,但是觉得很有创新,以后也可
: 以应用. 但是他不想去Google, M$当螺丝钉,想自己创业. 但是他不知道谁会买他写的
: library.
: 我自己做的项目跟他有类似,但是少一个CTO, 所以一拍即合. 我说我来替他找Business
: Model. 提了不少建议都给他否定了,因为他要做高技术,看不起普通的应用. 要有时间
: 研究,不能有Deadline. 不愿意跟不懂ML的人答扯,所以不能是用户接触高的项目.
: 最后想了我两个多星期,觉都没睡好,想到一个不错的方向, 几百亿的市场但是中小公司
| M****o
发帖数: 13571 | 6 所有的business都是卖产品或service，都需要source。这个人技术厉害就是你的
source，如果搞不定他就换一个source。如果搞不定source，再好的点子都是瞎掰呀。
加油！我看好你！;) | a***a
发帖数: 974 | 7 同意
【在 a*****1 的大作中提到】
: 找 Co-founder 是需要缘分的，呵呵
| p********r
发帖数: 1980 | 8 So many responses! So let me respond all in a single post. (Sorry for typing
in English, it is faster).
I don't think there is anything wrong with geek. Woz was a geek.
I actually like him for two reasons. 1) He has the basic requirement I have
for my co-founder: intellectual honest. It is OK to say that "I don't know".
Startup is full of unknown (risks), if there is no risk, it is a mature
company (in fact, no company has completely no risk, otherwise, its market
value should be infinity. I can prove that to you mathematically). And I had
my shares of elance Indian developers who are "experts" in "big data".
Enough said. 2) He is really into the ML community. For example, he is going
to write the core node in Clojure and knows Lisa does not have enough
library to build the code base (many people in ML field probably has never
heard about Clojure and Lisa).
A lot of ideas I have all have big technology risks. For example, a few
weeks ago, after reading the news, I have an idea of NLP-based quiz
generation Pearson Test Preparation software (reference paper: http://www.aclweb.org/anthology/P13-2043). It has huge technology risks. The idea in my original post has big technology risks as well, followed by distribution risks.
I am a product manager/system designer myself. So I can write or recruit
people to write basic ML codes. But if it fails (which is bound to happen
initially), I will have no idea whether it is that the technology is not
ready or I didn't use the right technology. So I need a co-founder that I
can trust who knows the technology inside-out. Therefore, I can tolerate a
lot of his geek-ness and even appreciate it somewhat, but if making money (
and making customers happy) is a distraction for him, then there is really
nothing I can do.
I also have one another idea that I am currently working on myself, which I
don't yet need co-founder because the risk is on the market side. That is,
it completely does not have any business model! LOL. But I am still pretty
passionate about it. After all, M$'s initial business model is to make
software for PC hobbyist, and GOOG's initial business model is...well, do
they even have a business model? Twitter certainly didn't have one...and you
can argue they haven't figured one that out yet.
It all depends on the projects and the biggest risks they have. I do know
one person who will be really good co-founders for me in the future...we
think in the same term in business/value perspective, and I completely trust
him as the best COO I'll ever know (even though he is still nobody). But, I
don't yet have any operation problem and all I have usually is technical
risks that only geeks can help me to address, for the moment. Until the next
risk item comes up.
【在 l******n 的大作中提到】
: 你找错人了
: 这种geek没有商业头脑，做不了战略伙伴。
:
: .
: Business
| l******n
发帖数: 9344 | 9 risk 和 CTO是两回事。
关键的是你们在主要问题上都没有共识，这还怎么搞？
typing
have
".
had
going
【在 p********r 的大作中提到】
: So many responses! So let me respond all in a single post. (Sorry for typing
: in English, it is faster).
: I don't think there is anything wrong with geek. Woz was a geek.
: I actually like him for two reasons. 1) He has the basic requirement I have
: for my co-founder: intellectual honest. It is OK to say that "I don't know".
: Startup is full of unknown (risks), if there is no risk, it is a mature
: company (in fact, no company has completely no risk, otherwise, its market
: value should be infinity. I can prove that to you mathematically). And I had
: my shares of elance Indian developers who are "experts" in "big data".
: Enough said. 2) He is really into the ML community. For example, he is going
| M****o
发帖数: 13571 | 10 不愧是搞marketing的，写东西不能简单明了些吗？呵呵 | s*f
发帖数: 807 | |
|
