Open AI strikes back: ChatGPT's new feature will end AI slop

I have a source at OpenAI that leaked a major feature they're going to put out, that will make up the ground they're losing to Claude. Here's an early draft of the press release.

***

OpenAI is rolling out a new feature with ChatGPT model 5.5 that provides special resistance to an ongoing issue with AI output: the telltale signs that the output is AI, which is a major turnoff and source of frustration on the internet.

The new feature, dubbed Preview Plus(tm), lets users go through ChatGPT's output and actually purge all the AI pain points.

• Em-dashes? Users can delete them.
• "It's not X but Y"? It's "X in Y clothing"? Users can replace it with their own authentic version a human would actually say.
• Hallucinated legal citation? Users can insert something from actual case law that they have personally validated applies to the matter in question.

Preview Plus(tm) currently only available to paid plan subscribers and users who have opted in to early feature testing.

OpenAI CEO Sam Atman is hopeful about the new direction Preview Plus(tm) will take ChatGPT. Earlier today, he told a press conference, "Preview Plus empowers users to actually stand between ChatGPT and the eventual consumer of its output artifacts. It provides a seamless way to merge human intelligence with cutting-edge AI, getting the best of both worlds."

OpenAI gave a screenshot of an early version of the feature that shows a user removing the dreaded em-dash. 

 

 

Leaked Google initiative: No more passwords!

I have an inside source that's claiming Google will be rolling out a new replacement for passwords and other secrets for authenticating users. They shared the upcoming blog post/press release with me. They're moving to a more "holistic" authentication system? Let's see if this pans out. In any case, here's the not-yet-released announcement.

***

Are you who you claim to be?

User logins protect websites from malicious actors, like spammers and trolls. So when you go online, only people with legitimate credentials can access the useful features of the site -- and others can't impersonate you. For years, you've used logins -- such as a username and password -- to prove to the site that you are who you claim to be, like this:

Some go even further and add a second factor to authenticate with, like an SMS code or one-time-password generator like you might have in the Google Authenticator app.

But, we figured it would be easier to just directly ask our users who they are -- so, we did! Following on our earlier success with No CAPTCHA reCAPTCHA, we’ve begun rolling out a new API that radically simplifies the login experience. We’re calling it "Credential-Free Authentication" and this is how it looks:

On websites using this new API, a significant number of users will be able to securely and easily verify their identities without (separately) having to provide credentials: no password, no rotating code. Instead, with just a single click, they’ll validate who they claim to be.

A brief history of user authentication

While the new login API may sound simple, there is a high degree of sophistication behind that modest interface. Authentication has long relied on attackers not having critical secrets, like a password or random number generator seed or other private information. You may have heard the traditional formulation, that authentication requires you to provide something you have, something you are, or something you know.

However, our research recently showed that it's about as likely for the genuine user to be missing the credentials as it is for a malicious actor. How many times have you forgotten your password or encountered a bug with your password manager? (Not GPM, of course!) Thus, challenging users for credentials is no longer a dependable test.

Furthermore, attackers are often able to steal user credentials, forcing providers to rely on a secondary layer of fraud identification, so as to lock accounts when users behave suspiciously. You've seen this if you've ever had a credit card declined for an unusually large or remote purchase.

Introducing Credential-Free Auhentication

That got our security engineers thinking: if we already have to analyze a user's behavior in order to catch account compromises, why not just use that as the authentication? It would cut two carrots with one knife! After all, an attacker might be able to guess your password or your credit card information, but they will never be able to mimic the full depth and breadth of how you interact with websites, from your browing history, to your cookies set, to the way you move your mouse.

Following the "No CAPTCHA" model above, we developed an Advanced User Analysis backend for logins that actively considers a user’s entire engagement with the the Internet to determine who that user is. This enables us to rely less on "Do you have the secret?" and, in turn, offer a better experience for users. Now, users can just click a radio button, and in most cases, they’re logged in. In fact, you'll rarely have to log in at all, because sites will "recognize" you, just like you don't have to show your ID to go into an event venue a second time if the bouncer recognizes you.

But are you really that person?

However, authentication challenges aren't going away just yet. In cases where our tracking cookies and other behavioral metrics can't confidently predict who someone is, we will prompt the user for additional information, increasing the number of security checkpoints to confirm who the user really is. For example, you might need to turn on your webcam or upload your operating system's recent logs to give a fuller picture.

Adopting the new API on your site

As more websites adopt the new API, more people will see Credential-Free Authentication. Early adopters, like Snapchat, WordPress, Twitch, and several others are already seeing great results with this new API. For example, in the last week, the number of support tickets for account resets on WordPress went down by 90%. Twitch reported similar figures -- and also was able to unmask several sockpuppets who had been manipulating discussions and vote totals.

To adopt the new CFA API for your website, visit our landing page for more.

Good users, we'll continue to keep the internet safe and easy to use. Bad users, it'll only get harder to hide yourselves and take over legitimate accounts -- sorry we're (still) not sorry.

***

Edit: Yes, this was an April Fools joke.

Some of my geeky tech jokes -- with explanations!

I know the line: explaining a joke is like dissecting a frog; you understand it better, but it dies. Still, not everyone will get these, and I figure I might as well have a place where you at least get a chance. So here are some of my own creations, explained.

---

Girl, you make me feel like a fraudulent prover in a stochastic interactive zero-knowledge proof protocol ... because I really wish I had access to your random private bits!

Explanation: In a stochastic zero-knowledge proof protocol, there is a prover and a verifier, where the former wants to convince the latter of something. But for proof to work, the verifier must give the prover unpredictable challenges. Think of it like a quiz in school -- it's not much of a quiz if you know the exact questions that will be on it.

The information to predict the challenges is known as the verifier's private random bits Those with a legit proof don't need this, but a fraudulent prover does. Thus, a fraudulent prover in a stochastic interactive zero-knolwedge proof protocol wants access to the verifier's "random private bits".

---

A historian, a geologist, and a cryptographer are searching for buried treasure. The historian brings expertise on practices used by treasure hiders, the geologist brings expertise on ideal digging places, and the cryptographer brings expertise on hidden messages.

Shortly after they start working together, the cryptographer announces, "I've found it!!"

The others are delighted: 'Where is it?'

The cryptographer says, "It's underground."

'Okay, but where underground?'

"It's somewhere underground!"

'But where specifically?'

"I don't know, but I know it's underground!"

'Slow down there. If all you know is that it's underground, then in what sense did you "find" anything? We're scarcely better off than when we started!'

"Give me a break! I just gave you an efficiently-computable distinguishing attack that separates the location of the treasure from the output of a random oracle. What more could you want?"

Explanation: In cryptography, an encryption scheme is considered broken if an attacker can find some pattern to the encrypted message -- i.e. they can identify telltale signs that it wasn't generated by a perfect random number generator, a "random oracle". Such a flaw would be called a "distinguishing attack". So in the cryptography world, they don't care if the attack actually allows you to decrypt the message; they stop as soon as they find non-randomness to the encrypted data. Applied to a treasure hunt, this means they would give up as soon as they conclude that the treasure location is non-random, which the cryptographer here things s/he's done simply by concluding that it's "underground".

---

So, 16-year-old Johnnie walked into an Amazon Web Services-run bar...

"Welcome," said the bartender. "What are you drinking?"

Johnnie replied, 'What've you got?'

"Well, we have a selection of wines and the beers you see right here on tap. But if you prefer, we also have club soda and some juices."

Johnnie thought, Wait a second. Why is he telling me about the wines and beers? Does he even realize ... ?

'Okay, I'll take the Guinness.'

"Bottle or draft?"

'Draft.'

"Alright, and how will you be paying?"

Johnnie only had large bills from his summer job and gave the bartender a C-note.

"Sorry, but I gotta check to make sure this is real." The bartender took out a pen and marked it, then counted out the change. Johnnie reached for the beer.

"Hold on a second! Make sure to use a coaster!" The bartender slipped one under the glass. "Okay, now enjoy!"

Johnnie lifted up the glass to drink. Before he was able to sip, the bartender swatted it out of his hand.

"WHAT ARE YOU THINKING!?! Don't you know 16-year-olds can't drink!"

Explanation: On the AWS site, they will gladly let you click on the "Launch server" button and go through numerous screens and last-minute checks to configure it, and only at the very last stage does it say, "oops, turns out you don't have permission to do that" -- so it's like a bartender that takes you through a entire transaction, even verifying irrelevant things (like whether the money is real), while knowing the whole time he can't sell to you.

---

How is a Mongo replica set like an Iowa voter?

In primary elections, they only vote for candidates they think are electable!

Explanation: Databases can have "replica sets" where there are multiple servers that try to have the same data; secondary servers depend on an agreed-upon "primary" to be the "real" source of data. Often times, the primary server goes down, so they have to decide on a new primary, known as a "primary election". But there are some restrictions on who they will vote for -- if they e.g. have reason to believe that a server can't be seen by other members, and in those cases it will regard that server as unelectable. So you can get funny messages about "server42 won't vote for server45 in primary election because it doesn't think it's electable".

Ending the tyranny of the mouse -- in web browsing

Since I plan to program professionally, I've ramped up my efforts to get by without a mouse, and I thought I'd share some key tools I've used to accomplish this.

For web browsing, the key is Pentadactyl, a Firefox extension that lets you do the things you want from the keyboard. (I would say all the things, but some websites are written so as to be unfriendly to it.) For example, if you want to click on a link, you hit f, and it pops up a keyboard code over every link, and you type the code to "click" on it. Here's what it looks like when you use it:



Other useful features are:
back/forward = shift+H / shift+L
page down/up = space bar / shift+space bar (these work without Pentadactyl)
half page down/up = ctrl + d / ctrl + u
search = / (yes, the slash key), then enter, then n/shift+N to search down/up
go to URL = o, space, [page url] (if you've entered something similar before, you can tab through the cached options)
go to URL in new tab = replace "o" in the above with "t"
open link in new tab: ;t , then it pops up hints as if you had pressed f but opens in a new window

However, you need to configure it a bit in order to get the most out of it. For example, as initially installed, it will remove your ability to use the familiar ctrl +c/v/a (for copy/paste/select all) due to its being based on the text editor Vim. Also, the hint keys (buttons used when creating a code that lets you click a click) are set by default to draw from the numbers 0-9, which are less comfortable to type every time you click on a link, and they are displayed too small to read.

To set your configurations, you need to create/edit a file called ".pentadactylrc" in your home directory. Here are the contents of mine, which fix the above problems:

"1.0rc1

loadplugins '.(js|penta)$'
group user
highlight Hint font: bold 10px "Droid Sans Mono", monospace !important; margin: -.2ex; padding: 0 0 0 1px; outline: 1px solid rgba(0, 0, 0, .5); background: rgba(255, 248, 231, .8); color: black; font-size: 14pt !important;
map -count -modes=i,n,v <C-c> <count><Pass>
map -count -modes=i,n,v <C-v> <count><Pass>
map -count -modes=i,n,v <C-a> <count><Pass>
map -count -modes=i <C-a> <Pass>
map -count -modes=i <C-x> <Pass>
set guioptions=bCrsmT
set hintkeys=asdfwervcxtgq
set hinttimeout=500

" vim: set ft=pentadactyl:

Enjoy your breaking the tyranny of the mouse! (And yes, I composed this entire post, including creating the link, without using the mouse. I may have mentioned that a few times before).

Mr. Ford, meet Boeing

You know how it's become a sort of cliche/folk-economics to say that "You should pay your workers enough so that they can buy the product you sell?" It's supposed to be what gave Henry Ford I his tremendous success with the Model T, and has become a staple of union bargaining.

For a recent example of this line of thought, here's none other than (former Secretary of Labor) Robert Reich arguing it, complete with reference to the Model T story.

Well, it recently occurred to me how underpaid I am. My employer modifies and sells large aircraft. No way can I afford that!!!

Did somebody say "raise"?

(This post made entirely without use of the mouse -- including for looking up and copying over links -- thanks to the use of the Firefox Pentadactyl extension. Give it a whirl!)

Addendum: To clarify, Boeing is not my employer, just a synecdoche for large aircraft manufacturers in general.

Well, I took the plunge and installed Ubuntu (again)

Three years ago, I tried to install Ubuntu. Let's just say it went so badly that I'm not even going to give the details for fear that someone I chewed out at the time will notice the similarity between my case and "that jerk on the Ubuntu help site three years ago" and target me for reprisals.

Needless to say, it actually worked this time, since I was smart enough to install it on a completely different computer as a hedge against failure (instead of merely trying to isolate it to a partition on a secondary hard drive) and because the development crew has gotten its act together.

And I have to confess, I enjoy it for the most part. There's still a lot to get used to, and a lot of settings to configure, but I was amazed how easy it was to get wifi working, to install Firefox plugins (note how I don't snidely call it "liarsux" anymore?) and how many Free (yes, "they" want you to capitalize it), useful programs come bundled, and yet the system has no bloat ... everything is fast. Unlike on Windows, there isn't a huge list of processes of questionable purpose running that you can't shut down.

I've also gone back to using Vimperator which I had blogged about before, which means that yes, I made this post without ever using the mouse. But of course, like with most user interface design, the half-genius Herr Stubenschrott had to ruin his own code's functionality. Previously, you would hit the 'f' key and a bunch of key commands would pop up over the links like "ds". Then, hitting "ds" would activate the link. But now, they're all numerical, like "24", which makes it much less convenient.

Stubenschrott, in his defense, now permits you to call up a link by typing the first few letters of it, which I had suggested allowing before ... but it kind of defeats the purpose when the key commands blot out the first two letters! And the entire link becomes highlighted and impossible to read! Fortunately, someone wrote a script that converts it back to the old way.

All in all, a seamless, enjoyable transition so far. Now, to move over the old hard drives, files, and email...

So I was right again. Now, let's fix inflation measures.

There's a story on CNN Money today about shrinking and degrading products in response to inflation. Unfortunately, it doesn't give more than passing mention to the real stickler in inflation, the "degrading" part, which is harder for measurers to notice.

Consumers are discovering more air in their bag of chips, fewer sheets of paper towels on the roll, thinner garbage bags and even smaller squares of toilet paper. (emphasis mine)

You don't say! I've been noticing this for a while, and haven't been convinced the BEA and BLS capture the impact. When you pay the same for a debased product, that is price inflation, and precisely what you need to measure. But like the fool who won't search for his keys outside of the light, the BEA and BLS don't do the lab testing necessary to incorporate critical quality-related aspects of products.

In my personal experience, I have noticed cereal boxes and paper cups as being flimsier and thus harder to hold -- about as big an inconvenience as you can tag onto such a simple, trivial product. Soda bottles also had confoundingly irritating changes: in addition to the 25% vending machine price increase, they shrunk the cap height beyond all reason so that it's nearly impossible to get a good enough grip to twist open with your hands. The fact that Coca-Cola even made this decision is a testimony to either a) the low quality of their engineering teams, or b) how desperately they needed to debase the product. Neither is encouraging. (To their credit, the caps have returned to "good enough", meaning they've hidden the price increase somewhere else.)

I should feel fortunate to live in a country where "difficulty in opening products" ranks highly enough to complain about. But that's also worrying: in a country with such enormous, overflowing wealth (which the US has, right?) shouldn't producers have kept such noticeable inconveniences out as a matter of course? Something's not right about that picture...

So, if you really want to measure inflation, you're going to have to track these very tricky quality changes. But there's an alternative: focus on measures were this quality debasement just isn't possible. As I'm sure I've argued here and on several boards by now, the ideal candidate is an insulin index which does the work of policing quality improvements for you. If you debase insulin, someone dies. The other benefits are:

-Steady, predictable demand
-Global market with many buyers
-Many inputs, so it's immune to any one specific input's volatility
-No transient intellectual property effects

Which probably accounts for why such information is so durn hard to find!

Second, in addition to capturing quality degradations, they need to fundamentally rework how luxury-type items are accounted for. Those typically "scale" with what other people have. Faster computers mean enabling nicer software, but they can also mean having to pay for hardware I don't need, as the older stuff isn't available, and my current one can't run the latest software that assumes I have a faster machine. And the value I can squeeze out of it doesn't increase one-to-one with the MegaHertz rating!

I absolutely accept that modern technologies have vastly expanded the entertainment and learning options available to me, but an inflation measure must at the same time account for when food and energy prices put the squeeze on me.

I'd be interested in transforming these ideas into an academic paper, except there are a few things ahead on that list...

Setting SYNC Straight (Or, SYNC SUCX)

So I thought I'd try out "SYNC" by Microsoft and Ford. It's the system that lets you control your cell phone and iPod by voice. Hey, great stuff, right? What a convenient way to bounce around between your playlists, your artists, your podcasts. And even better: after having driven around in a car with merely a one-way input jack, I certainly realized the value of having an mp3 interface that a) is hidden from view so it doesn't have to be reconnected each time, b) charges as you drive, and c) will pick up right where you left off when you come back to your car.

Sync fails about as badly as possible while still having the general description given to it.

Let's go over how:

1) Sync's marketers make no attempt to ensure that your first experience with Sync is with a qualified professional, trained in the system. While there may be some dealers that provide for such (NOTE edited to clarify), Ford and Microsoft don't seem to be interested in making that a requirement to sell their hypercool, "let's make up for previous failures" product. In the two dealerships I went to to check it out, neither one knew how to operate it. On the first one, the salesman made it even worse by trying to figure it out for the first time when he tried to show it to me, even while knowing 24 hours in advance that I'd be there!

2) The Sync site is syncularly useless. No "set an appointment to try it out". No "list of dealers that offer". (Just a list of cars on which it's offered, and dealers that sell those cars ... who may or may not have Sync cars.) The list of supported phones? You get to download a pdf for this basic list. It has no forums, no feedback. Each part of the site takes a long time to load. It has no instruction manual that can tell you exactly how to do what you want to do with it (and therefore IF) you can.

3) The claimed podcast support is a LIE. While being allowed to *play* your podcast is supported, anything beyond that is little more than an afterthought. You can't tell it to play a specific podcast, and one post I found says that, as best he could discern, you have to figure out the artist, and tell it to play that. Oh, and good luck maneuvering to the specific track within that podcast that you want to hear. Hope you figured out the number!

And telling it to "start from earliest unlistened podcast and go forward in time"? HAHAHAHAHA! You're kidding, right?

4) The Ford Focuses that I looked at don't conceal your iPod from view at its point of input (as would make sense to protect it from thieves that look into your car to find stuff to steal). Okay, okay, maybe a design oversight. But no, they actually advertise how they have a special compartment to easily hold your iPod [link forthcoming], and even though the designers though to have such a nice iPod holder, it somehow never occured to them to put some kind of cover on it to keep it out of view!

5) Despite what I wanted in c) above, it doesn't just pick up where you left off each time you turn the car back on, and doesn't even bother to stop playing when you turn off the ignition. Okay guys, I can understand how neato it is to keep stuff running when you turn off the engine so you can keep up the mood when seducing your date. But what about the other 30% of my life when I'm going on various errands while I want to hear my podcasts? I want to turn off the ignition, leave, come back, start it again, and continue where I left off, all without having to fiddle with my iPod. Ease of use WAS the whole point, REMEMBER?

What's worse, as others have pointed out, the voice will nag you about your lacking metadata every time you turn your engine on! That's on TOP of you having to tell it to go to your USB input every time you start the car.

6) They want you to use Sync for a whole bunch of functions -- it's a cornerstone of their marketing (see the top ten the bottom left) -- which would imply you're going to be hitting the "voice input" button quite a bit. Yet I look at the steering wheel and find that ultra-frequently-used button on the harder-to-reach inner row of the steering wheel controls.


(picture from Yahoo Autos; annotations are mine; click to enlarge)

HUH? Did your designers miss the day where they taught you make the most frequently used stuff, the easiest to access?

7) It doesn't automatically shuffle your music playlists, and even after I told it to, it STILL started on the same track each time I called up the playlist. No, I'm not going to use your ultra-cool "play more often/less often" function to get this kind of thing not to happen.

8) When I hit the phone button while driving -- an error, since I had no phone set up with it nor had I turned on my own phone's bluetooth functionality -- it adamantly REFUSED to stop searching for a phone and go back to playing other media (radio or mp3s) on the car's sound system. No matter what button I hit -- the change media button, the phone button, the radio button, etc etc etc, I could not start playing the radio or other mustic. Do you understand that? Because I hit one button that I didn't really want, that caused the entire audio system to become useless until I turned off the engine. WHAT? Did your designers miss the day when they talked about containing failures? Hitting one button in error should not turn off the car's sound permanently.

Well, at least they got the voice recognition by and large correct, as well as the charging of the iPod. (Wait, I didn't check the last one there ... and frankly, I can't give them the benefit of the doubt at this point.)

What is wrong with you designers at Ford and Microsoft??? I could design a better system in an afternoon. Heck, I just did. Implement these suggestions and you have a product worth buying.

Oh, and in case you guys get your act together and are wondering: my salary requirements are $150k/year plus relocation expenses and benefits, plus, in the case of Ford, wage payment priority over pensioners. Y'all have my email.

A few more giant steps toward giving up my mouse

Many of you may not know I've been in a quest to abandon use of the computer mouse entirely and do everything from the keyboard, as part of my broader goal of improving ease of computer use, which itself is ... nevermind, I'll cut it off right there.

The point is, I've found an absolutely marvelous tool that lets me use the keyboard for a notoriously mouse-intensive task -- web browsing! With this tool, Vimperator (a Firefox plugin and apparently modeled after the Vim text editor, which may or may not have something to do with the "vi" environment), you can "click" on any link just by hitting "f", which brings up a key command for each link on the screen, and then typing that command! While it has many other features (that I'm still slowly learning), that by itself justifies installing it.

In fact, I even created this entire post without using the mouse, and yes, that includes logging in and copying the link from another tab.

With such an amazing tool, how can I criticize it? Like this:

The time for the link key commands to pop up is too long for my taste and makes it not time-competitive with the mouse. (This is probably not easily improvable -- the programmer is no fool!) Also, it doesn't have enough alternate commands to make one-handed use very feasible. (Yes, you just thought of a naughty comment and no, we don't care.) ADDENDUM: oh, and how the link key commands ("hints") don't seem to exploit the letters that already exist in the links (as native alt commands do), plus they overlap the links without any transparency features that let you continue to see the text of the link while the hint is on top.

Still, compared to all that vimperator makes possible, these are tiny nitpicks. Hats off to Herrn Martin Stubenschrott!