Learning is not compulsory. Neither is survival.
W. Edwards Deming
I am doing a startup!
Cross-browser testing from your browser!
I have written my fourth book!
Be faster than Larry Wall in the shell!
You're viewing a comment by T. Human and its responses.
BTW, I read the article mentioned by a poster above: "AI has been solved". Might I politely say that instead of writing about how you've been oppressed, a short clear description of what your idea is would speak volumes: but an impressive demo would be indisputable.
Computer science is unlike other technical fields. If I write a theoretical paper about astrophysics, I'm not expected to go out and detect the phenomenon I'm describing: but computer science is so young that theory lags behind practice in many cases.
If you have something new, show it to us. It can be a toy demo, but it should do one new thing, or demonstrate clearly one new technique.
To be honest, just the formulation is unconvincing. There is no "problem of AI" - there are hundreds or thousands. Some of them we've already almost solved (like voice recognition/speech to text) - some of them we have made little progress in 50 years (like reasoning about facts taken from text).
Your Mentifex model has some aspects of interest: it reminds me to some extent of Hawkins' writings in "On Intelligence", which I commend to your attention. And he doesn't have a demo either. :-D
But he does have a lot of corroborating information from neuroscience, which is fairly convincing. And he has the monster track record. :-D
My suggestion if you want to get read:
1. build a demo; and/or
2. find some actual experiments that corroborate your results, or even:
3. propose an experiment where your theory makes specific predictions that other theories do not.
(and if you're trying to publish in scientific journals, lose the fancy name - it sounds like "TimeCube" or something - give your paper a descriptive name without advertising claims like "A multi-level model of the human brain.")
(why do I need your e-mail?)
It would be nice if you left your e-mail address. Sometimes I want to send a private message, or just thank for the great comment. Having your e-mail really helps.
I will never ever spam you.
(Your twitter handle, if you have one.)
* use <pre>...</pre> to insert a plain code snippet.
* use <pre lang="lang">...</pre> to insert a syntax highlighted code snippet.
For example, <pre lang="python">...</pre> will insert Python highlighted code.
* use <code>...</code> to highlight a variable or a single shell command.
* use <a href="url" nospam>title</a> to insert links.
<a href="url" nospam>title</a>
* use other HTML tags, such as, <b>, <i>, <blockquote>, <sup>, <sub> for text formatting.
Type the word "rocket_101": (just to make sure you're a human)
Please preview the comment before submitting to make sure it's OK.
Peter Krumins' blog about programming, hacking, software reuse, software ideas, computer security, browserling, google and technology.
Reach me at:
Or meet me on:
Subscribe through an RSS feed:
Subscribe through email:
Enter your email address:
Delivered by FeedBurner
See all top articles
See all downloads
See more detailed list of recent articles
See more detailed category information
See more detailed list of all articles