Noah

In the band, great food, video games, and the technology that runs the world.

[Noah @horn978]

This thread was well worth the read: starting a company from $100, just as the AI would do it. I’ve read and seen all sorts of entrepreneurial material, and HustleGPT is pulling a lot of the right shots to start a business. Of course, the human counterpart is pretty good too

[Jackson Greathouse Fall @JazzFall]

I gave GPT-4 a budget of $100 and told it to make as much money as possible.

I'm acting as its human liaison, buying anything it says to.

Do you think it'll be able to make smart investments and build an online business?

Follow along 👀

[Eliezer Yudkowsky ⏹️ @ESYudkowsky]

I don't think people realize what a big deal it is that Stanford retrained a LLaMA model, into an instruction-following form, by **cheaply** fine-tuning it on inputs and outputs **from text-davinci-003**.

It means: If you allow any sufficiently wide-ranging access to your AI model, even by paid API, you're giving away your business crown jewels to competitors that can then nearly-clone your model without all the hard work you did to build up your own fine-tuning dataset. If you successfully enforce a restriction against commercializing an imitation trained on your I/O - a legal prospect that's never been tested, at this point - that means the competing checkpoints go up on bittorrent.

I'm not sure I can convey how much this is a brand new idiom of AI as a technology. Let's put it this way:

If you put a lot of work into tweaking the mask of the shoggoth, but then expose your masked shoggoth's API - or possibly just let anyone build up a big-enough database of Qs and As from your shoggoth - then anybody who's brute-forced a *core* *unmasked* shoggoth can gesture to *your* shoggoth and say to *their* shoggoth "look like that one", and poof you no longer have a competitive moat.

It's like the thing where if you let an unscrupulous potential competitor get a glimpse of your factory floor, they'll suddenly start producing a similar good - except that they just need a glimpse of the *inputs and outputs* of your factory. Because the kind of good you're producing is a kind of pseudointelligent gloop that gets sculpted; and it costs money and a simple process to produce the gloop, and separately more money and a complicated process to sculpt the gloop; but the raw gloop has enough pseudointelligence that it can stare at other gloop and imitate it.

In other words: The AI companies that make profits will be ones that either have a competitive moat not based on the capabilities of their model, OR those which don't expose the underlying inputs and outputs of their model to customers, OR can successfully sue any competitor that engages in shoggoth mask cloning.

[Noah @horn978]

@foxfortmobile To be fair, it is a technical security flaw in which a hacker could downgrade your device to a more vulnerable OS version. But still a much rarer scenario and it sucks for the jailbreaking scene

[AltStore.io @altstoreio]

The new “CheatBase” allows you to easily browse and enable known cheats for games 😈

For now it's limited to select DS games, but we hope to add more games & systems in the future!

1. Open "Cheat Codes" from the pause menu
2. Press "+"
3. Choose “Search CheatBase”

[Noah @horn978]

@altstoreio If you want to expedite the process of getting more cheats added, check out the GitHub: https://t.co/xmhRTex5CY

More great things are also on the way!

[Provenance-Emu 🎮🕹 @provenanceapp]

You won’t believe how scary good #chatgpt is at converting #objc to #swiftlang

It replaced cumbersome legacy UIKit to trimmed swift with CG* helpers, proper weak and opt, throws, string based auto layout, finding delegate protocols in files not provided. https://t.co/YH8cN1uxgX

[Noah @horn978]

The last year progress on Twitter 😭

[Years Progress @YearsProgress]

▓▓░░░░░░░░░░░░░ 11%

[AltStore.io @altstoreio]

Delta 1.4b5 is now available for all Patrons!

This beta includes several great new features:

👾 Integrated ‘CheatBase’ (thanks @horn978!)
🔕 “Respect Silent Mode” setting
🎶 Auto-mute game audio when music is playing

Read more on our Patreon: https://t.co/tGXEsedI2R

[National Park Service @NatlParkService]

You know it’s cold outside when you go outside and it’s cold.

[9to5Mac @9to5mac]

AltStore creator explains what third-party app stores could mean for the iPhone https://t.co/IhkBPxTbpr by @filipeesposito

[Noah @horn978]

We are getting real close to the ability of language models being able to produce and run their own code. Is this how true AI is going to be born?

[OpenAI @OpenAI]

Our new embedding model is significantly more capable at language processing and code tasks, cost effective, and simpler to use. https://t.co/Wyba0DZGBF

[Noah @horn978]

@rileytestut I foresee a system where developers / publishers can release their apps outside of the App Store. But the only way to “install” from this third party App Store, is if the app is signed with a special certificate which happens to be sold by apple for a pretty penny.

[Noah @horn978]

@trakt Thank you for the updates, your work is appreciated!

[Noah @horn978]

@rileytestut @tapbots Moving AltStore tweets over there too, or just personal dev thoughts for now?

[Noah @horn978]

@Affordablecodot @therabbithole @elonmusk No, it’s because you hate half of the opinions that exist 🤷🏻‍♂️

[Noah @horn978]

@rileytestut I really look at those accounts and pray that they are bots, because I want to retain my belief in humanity and what we choose to spend our free time typing…

[Noah @horn978]

And to do that without human intervention, will be the pinnacle of AI.

[Noah @horn978]

One of the most troublesome facets of AI programming (specifically with LLMs), is the reliance on human literature in vast quantities. Why, you might ask? Because the internet is not a source of truth, which makes your model more fiction than fact: https://t.co/n6jx4STt3g

[Noah @horn978]

There are definitely a lot of interesting methods to improve from here, but I think decreasing the reliance of models on vast datasets for training, is the ultimate goal. At the very least, the LLM needs a way to decipher what’s good training and what’s bad training.

[WeRateDogs @dog_rates]

a few of my favorite memories