_fol8

Joined 9 years ago

Karma

187

Submissions

359

Member since

May 6

View on Hacker News

Recent submissions

Comment · 3 years ago

The major journals have absolutely no accountability. In any other market, if the product doesn't work or harms someone the company goes out of business or the maker is sued. Not so in journals. So, why do we accept it? Because there's no other way for the layman to determine what makes a good professor, because by definition, they are smarter than us (or at least they're supposed to be), and so we (the general public) are not able to tell if they are good at what they do or not.

So - the answer we have is peer review, which is just the foxes guarding the hen house. There's no other solution that's been proposed that makes any sense in a self reinforcing market manner. Having some post-docs suddenly become concerned about this and hire a bunch of undergraduates to start using to comb excel with spreadsheets will be useful until everyone loses interest. The price of a can of Coca-Cola isn't useful until people lose interest - it's market priced by millions of customers at every minute of every day.

Until there's a solution to this problem that makes sense this will keep happening over and over again.

Comment · 3 years ago

This sounds like a puff piece and it's hard to be all that excited about the products that are being touted in the piece -

He decided instead to quit and pursue his idea to start Feasier, a platform that aggregates home furnishing listings from different stores into one place.

So...what's the moat? Why can't anyone build a clone sight and come in with slightly cheaper shipping and undercut his business?

Or this -

Zhu says she is working every day on Maida AI, which automates health care administration tasks like patient intake and note-taking.

From what I know of the current health care tech situation, everything uses EPIC which is terrible technology, but it's entrenched because it follows regulation and probably pays off some senators/regulators/insurance companies - much like every DOD government project. So unless her product can follow every regulation as well as EPIC (it can't) she can't win in an entrenched market that has a moat she can't cross.

I don't doubt there are a bunch of unemployed web developers out there (I'm one), and a lot of VC money that doesn't know what to do after crypto imploded, but these are uninspiring. Of the thousands of web dev companies maybe one will become popular, but it would be from random chance. This just sounds like a bunch of unemployed and desperate people buying lottery tickets, but with computers.

The people who are going to make money are the people who are able to corner the energy market or the server and computation market. The innovations will be from a few brilliant academics who make the next transformer model or the next quantum computing model and so forth, which makes up a few thousand people. The "democratization" of the internet just now means everyone is equally broke. Shit, you can make websites now by drawing them and then having an AI make the code (as edge-case clunky as that is). So anyone with a Bamboo pad and a "million dollar" idea can draw one up and buy a web address.

So what?

Comment · 3 years ago

I read a NY Times article by Ted Chiang today in which he made a kind of "stochastic parrot" argument for chatGPT - https://www.newyorker.com/tech/annals-of-technology/chatgpt-.... I believe, on the other hand that chatGPT may eventually be able to generate AGI, but that this will occur emergently and spontaneously. In other words, it will be difficult to predict.

One of the conditions for this, is for chatGPT models to start being able to write their own code in order to produce models of themselves that are more accurate and more efficient. Given this ability, and some fitness criteria, genetic algorithms may be used to create new LLMs. This sounds like science fiction, but once the compute requirements come down for these models (by a couple orders of magnitude), I believe this may be possible.

To what extent does your model allow for semantic models to create semantic models that are themselves more efficient in relation to some fitness criteria? Can I tell a model "You (model) I want you to reproduce using interaction with these other models (some collection of other models) and have the child model offspring be more efficient according to this criteria [for example the resultant models will create short stories that are more likely to receive high ratings on a subreddit devoted to short stories]".

You would need to get around the "model pollution" problem in which LLM models pollute the space for which the models generate data because other models are producing web artifacts (Ted Chiang's Xerox of a Xerox problem). I call this the problem of alpha (direct experience). One of the ways I've thought of to fix this is to have models trained on direct user input (such as cell phone video and pictures from a single user) - I have to admit that I got this idea from Neal Stephenson's Snow Crash (see Gargoyle). If your platform can integrate with visual processing this may have a high information density - object detection in daily videos demonstrating how objects are related to each other in the real world of the user and correlating these into a semantic network.

I'd also suggest that Obsidian integration might be useful.

This is exciting, thanks for making the Fixie SDK public.

Comment · 3 years ago

I'll believe aliens exist when little green men walk up to me, shake my hand, and start discussing the engineering particulars of faster than light travel. Until then...