Everybody keeps talking about post quantum encryption. We should also be talking about post AI web browsing.
@jerry but talking about post-AI browsing makes me weep
@jerry Aka "Make Web Browsing Great Again"

@jerry My strategy: Webfeeds, & being more strategic about my search engine use. Knowing SearchMySite, DDG, Stract, & Wikipedia's strengths & weaknesses.

It's serving me well for now!

@alcinnz @jerry good list. check out kagi/mojeek instead of ddg (kagi also has a small web search feature)

@breadcat @jerry I might!

I do need to keep my options open...

@breadcat @alcinnz @jerry I’ll def check them out. DDG has been my default for a while now but lately the results have been less than satisfactory for some reason. I end up going back to Google more often lately and that’s not desirable.
@jerry remember blog rolls? i miss those simple, curated lists of links to other people's interests 
@jerry
100%. It seems like we have far more of a plan to deal with PQC than AI web browsing.
@chillybot @jerry One of those things is part of another bubble, one is not.

@jerry Everyone will be assigned a topic via lottery. You will be required to curate a short list of URLs that are NI-produced resources for it. Moderation and metamoderation will produce an attestation network, such that to search, you ask everyone you know "Hey, does anyone know how to do dovetail joints?" and tag it with carpentry woodworking dovetail. Then it finds your grandpa's ham radio buddy who is also a cabinetmaker, who calls you up on the phone.

...Okay, maybe that won't work.

@log @jerry That was pretty much the Yahoo approach — full curation — and it went fine until the volume of information available on the web becoming overwhelming for the moderators. Also, reasonable people may disagree on the quality of information. That’s a good thing. Dispute is *required* for knowledge to advance.
@jerry As much as I love AI, people really need to take ownership of its output before posting it. Proofreading, editing, and revision are all needed still.

@wagesj45 @jerry AI is only half of the problem. The other half is SEO and together they produce the poisonous mixture, that is destroying the web.

Owning the content was never and will never be on the agenda, as long as it brings traffic.

@jerry
Who wouldn't want a web where everything is confidently wrong?

Google Classic: Here's the result based on a unit conversion formula.
New Google: Here's a statement that we believe sounds plausible.

Progress!

@wdormann @jerry
Isn't this short vs. long billion?
@jerry I have a bad feeling it will mostly happen on either archived copies of pre-LLM sites, or an alternative network.
@jerry please Jerry, one really scary thing at a time.
@jerry voight kampff, but for search results, resumes, academic resources, and blogs 💯💯💯

@hacks4pancakes @jerry
I know it's tongue in cheek.

Detecting AI is an interesting problem.
So far all the "detect AI" systems are appallingly bad and some likely use AI to do that (!).

Even assuming you can determine #AI algorithmically, with no need for AI. AI engines would just use the algorithm NOT to look like AI.

Back in the day when the status of AI was not clarified in academia (it became permitted in my course). I would ask the engine to introduce punctuation, capitalisation and 3 spelling errors per page and use vernacular English.
#promptengineering #AIClassroom

@hacks4pancakes @jerry A basilisk for AI at the top of ever human created webpage?

@hacks4pancakes @jerry I've been seeing such a flood of AI crap, and so many posts which really look like AI training queries.

But you adding the Voight Kampf test just makes it hit the anxiety harder.

@hacks4pancakes @jerry I think of voight kampff almost daily as I go about my day job. A completely new development as of this year.
@jerry please don't forget about Copilot+Toilet
@jerry but will there be any post-AI browsing? 🤔
@jerry not now I'm still setting up AI in a Box
@jerry I thought web browsing was already a lost cause? Any searching for anything is already 95% Blackhat SEO scam sites/blog farms / or AI garbage images or pushed product advertisements. For all intents and purposes web browsing is already dead.
@jerry it's getting so bad we might have to go back to finding information in books and then remembering it somehow. I'm no longer cut out for this.
@jerry I’m looking forward to post VC hype IT generally. One day? It’s not viral growth or blockchain or AI specifically, it’s the ecosystem around these things that breeds them that I want to move past, because even after the AI craze dies down, there will always be a new stupid fad with pump and dumpsters behind it.
@jerry yeah this is becoming very important, and frankly needs to be solved well before we worry about encryption being broken, as that is still in the future rather than the present

@jerry In 1992 a coin was termed by the PGP creator himself, Phill Zimmerman: web of trust.

Make public PGP signature based identities cool again? Create robust networks of trust where users sign each others keys. There are levels of trust, the most of which being somebody you personal know and have signed (ideally IN PERSON).

We build out networks of good faith actors. And when some actors are deemed malicious, democratically from a strong network, the network can dispel them. Or lower trust.

@jerry @hacks4pancakes especially as it's essentially about rsa-22...
@jerry @pitrh Earlier today on a call with a coworker I was trying to get to the GitHub asset upload/download actions - first search it linked me to an abandoned library that wasn't the official one, and then a bunch of blog posts. The actual link was behind the infinite scroll.

@jerry

Neal Stephenson predicted this in Termination Shock. He also predicted some folks being able to curate versus not, and extrapolated those effects on society.

It was not an optimistic book.