Introducing the First-Ever AI Journal & Podcast Made By AI | by Shaked Zychlinski 🎗️ | The Startup | Apr, 2024

An Exploration into Automated Information Era with No Human Oversight

Tech by AI is offered at techbyai.information. Open supply code is offered on github.com/shakedzy/techbyai

Tech and AI are advancing quick. Actually quick. So quick I can’t sustain with the tempo, and located myself misplaced when attempting to. There are new discoveries and fashions on a each day — generally hourly — foundation, a lot information to devour, so many tweets to learn, how do I make all of it work?

Wouldn’t it’s nice if somebody — or let’s say, one thing — would collect all of the information for me, filter out solely the issues that basically matter, and summarize them, so I can get all of the information with morning espresso?

So I’ve determined to perform a little experiment — a social experiment with no people concerned — and easily let the generative fashions learn, mixture, filter and summarize the vital information for me. Every part shall be finished routinely, with none human intervention. How good will the end result be? Will it is smart? How a lot will it value? There’s just one option to discover out.

Selecting a Mannequin

Clearly, probably the most essential half is which LLM to make use of? There are such a lot of available on the market — with new ones becoming a member of each day — this isn’t a trivial name. I noticed I’ve two predominant necessities from the LLM I’ll select:

1. It wants a protracted context window. The mannequin will scan via and browse a number of totally different articles earlier than serving me with one thing, so it wants the flexibility to retailer a number of information in its reminiscence.
2. It must work properly with exterior instruments. Clearly, the mannequin shall be required to go looking the net and entry web sites on my behalf, so working with exterior instruments in an efficient manner is essential.

With these two necessities in thoughts, I got here to the conclusion that GPT-4 Turbo is the mannequin to go along with. So now that I’ve the mannequin to energy my newsroom, it was time to ask how will the newsroom function? Am I simply going to ask GPT to “summarize information on the internet” for me, or do I would like it to work together with different folks — or fashions — like an actual newsroom?

Brokers

Impressed a lot by Microsoft’s AutoGen (although I haven’t used it on this venture), I’ve determined to go along with the second possibility — I’ll have a number of brokers, every with their very own position, interacting with each other to create a each day concern for my AI information journal. After some trial-and-error, I’ve converged to 4 forms of brokers, working collectively:

1. Editor-in-Chief. That’s the agent that governs every part, and finally has the final phrase. The Editor doesn’t write any article — they solely edit the reporters articles. The Editor can be the one to temporary the reporters about what to search for, and in addition has the ultimate choice in what shall be featured within the each day concern.
2. Reporters. Reporters are the brokers which do the analysis on-line, decide the highest articles and write about these chosen by the Editor. There’s multiple reporter, as the purpose is to have every which a unique system immediate, which ought to ideally end in totally different web-searches and totally different article choice.
3. Educational Reporter. One of many issues I shortly realized is rather like people, giving brokers to many choices yields confusion. As a substitute of asking the identical reporters to do analysis each on-line and on Arxiv, I break up the duties, and gave the academic-research job to a separate reporter, dealing solely with this.
4. Twitter Analyst. Within the discipline of AI, information and developments generally begin off as tweets earlier than getting headlines on extra conventional media. Realizing that, I created an agent specializing in looking information on Twitter, which then notifies the editor what everybody’s is speaking about.

Having established these roles, it grew to become clear that I must focus now on offering them with strong instruments to successfully collect and course of data. This requirement led me to discover and arrange the mandatory digital infrastructure.

Instruments

Speaking with the outside-world is a very powerful factor for my newsroom brokers to efficiently accomplish their assignments. Listed below are the instruments I wanted, and the way I created them:

1. Internet Search. The standard of the journal will instantly correlated to the brokers search capability. Subsequently, I gave them entry to Google Search. Getting began with entails establishing a Google Console account with an lively Search API, and establishing a Customized Search Engine. As soon as finished, the official Python bundle might be put in from PyPI: google-api-python-client. The documentation isn’t nice, although.
   (FYI, there’s one other free, out-of-the-box, no-questions-asked possibility by DuckDuckGo).
2. Accessing Web sites. As soon as discovered, the articles must be learn. In Python, making a easy too to scrape textual content from web site might be finished with just a few traces of code utilizing requests and BeautifulSoup.
3. Accessing Arxiv. A little bit documentation-lacking too, however Arxiv makes it very straightforward to go looking and obtain PDFs from it. There’s additionally a fairly straightforward to make use of Python library named arxiv. We’ll want one other library for parsing the PDF recordsdata. I used PyPDF.
4. Accessing Twitter. This one is somewhat difficult. Twitter beneath Elon Musk fees $100/month to entry Twitter API. As a workaround, I used Google search whereas limiting it with website:twitter.com. This appears to be working fairly properly for public tweets, that are the overwhelming majority.
5. Journal Archive. Information can someday be duplicated, and a subject mentioned on one website at present might need appeared on one other yesterday. I wished to provide the Editor an choice to seek for articles within the journal’s archive, and examine if there are any related headlines from earlier than. To get this finished, I created embeddings of every article within the journal, and permit the the Editor to go looking in the same option to how RAG works. As this little or no information, I used a naive Numpy array and Pandas DataFrame because the vector DB.

With the instruments in place, from net search capabilities to Twitter information entry, I used to be able to outline the each day operations of my AI-driven newsroom. This setup dictated how the brokers would work together and the way the complete course of would unfold every day.

The Routine

Now we’ve got the decided the brokers and arrange their instruments, it’s time to find out how the each day routine will appear like. I had two conflicting pointers right here — the primary was to let the brokers work together as a lot as wanted with each other, and the second was to restrict their interactions in an effort to cut back prices. Ultimately, the next routine was the one which labored greatest for me:

It goes like this:

1. The routine begins with the Editor getting a common overview of what I’m anticipating of the journal to be — what’s the fields and particular matters I’m it.
2. Within the meantime, the Twitter Analyst comes up with an inventory of individuals to observe on Twitter, and checks what they’re speaking about. It compiles an inventory of developments, and sends them to the Editor.
3. The Editor takes into consideration all these inputs, and creates a briefing for the reporters, asking them what to search for and write about.
4. The reporters search around the net and Arxiv, and ship an inventory of the perfect gadgets they discovered again to the Editor. Who decides what are the highest gadgets? The reporters themselves, after all.
5. The Editor seems in any respect the solutions and does a number of issues:
   – It decides what are the gadgets to be featured within the concern, and asks the reporters to jot down
   – It combines a number of solutions about the identical matter from totally different sources, to keep away from duplications
   – It seems up the articles matters within the Journal Archive, verifying this matter wasn’t coated already
6. Reporters summarize the articles, and hand their drafts to the Editor.
7. The Editor has the ultimate say, and has the choice to edit the texts. The ultimate edit is being served to me

This whole course of takes rather less than 5 minutes, and prices fluctuate from $1 to $5, relying on the size of texts learn by the brokers.

After outlining the each day routine that effectively makes use of our brokers and instruments, I centered subsequent on the distinctiveness of every publication. This uniqueness is primarily pushed by the system prompts of every agent, curated to inject selection and depth into the content material they generate. Which is why I made a decision I gained’t be the one writing them.

Because the Editor is the one in cost, the primary job it will get is to rent the reporters. The Editor is requested to explain the traits of the reporters which would be the greatest match for the newsroom. I ask the Editor to explain them in second physique, as if addressing them instantly, telling them who they’re. I then take these descriptions and use them because the reporters system prompts.

And who decides what’s the system immediate of the Editor? For that I take advantage of one other agent, with just one job — to explain to me a number of totally different editors and their traits, once more in second physique. From these I randomly decide one, and assign it because the Editor. Add to that the truth that all brokers temperature is ready to ~0.5, and also you’ll notice that for those who run the identical routine 10 instances in arow, you’ll get utterly totally different points. Each concern is exclusive.

Creating content material is nice, nevertheless it must be served one way or the other. I made a decision to go along with a easy and environment friendly answer — GiHub Pages. All I wanted to do is to ensure the ultimate edit is written in Markdown. I used a clear and MIT-licensed Jekyll theme I discovered on-line, and that’s just about it — I received an internet site. I additionally built-in GitHub Actions to set off the routine each morning, so when wake there’s a brand new contemporary concern prepared for me.

However then I noticed that I really wish to get my information after I stroll my canine within the morning — and it’ll be nice if the information might be narrated for me. So I added one final section to the routine — narration. To maintain it easy, as I’m already utilizing OpenAI API each for GPT and the embeddings, I made a decision to make use of the corporate’s text-to-speech API too. And as Jekyll and GitHub Pages render my web site each time a brand new concern is added, creating an RSS feed is simple. Now, for those who didn’t know, apparently establishing a podcast solely requires one factor — an RSS feed. So, in a matter of minutes, my information narration grew to become obtainable on Spotify, and now I get me information each morning whereas I’m out for a stroll.

Whereas the each day prices have been at all times within the vary of $1 to $5, as days glided by, I seen they stabilized round ~$3.5. Which is isn’t rather a lot, however that’s nonetheless greater than I used to be anticipating, because it provides as much as ~$105 a month. So I took a deeper look into the prices breakdown, and seen that the analysis section — the one the place the reporters search on-line for articles — was the most costly a part of the method, reaching ~$2.7. Is there a option to cut back the prices with out affecting outcomes? Sure — decreasing tokens.

Whereas English phrases are generally both a single token or two, URLs are a bit extra problematic. As there aren’t any areas, and phrases are both separated by dashes, slashes or by nothing in any respect, and are sometimes blended with numbers — and are additionally often very lengthy — I noticed a single URL may require even 27 tokens. Think about the quantity of URLs which are being processed — that turns into a number of tokens.

The answer was to map URLs to IDs. Behind the scenes I changed all URLs with a numeric ID, and gave that ID to the brokers. My code transformed URLs to IDs and vice-versa. I selected numeric IDs for a purpose — all numbers which have as much as three digits (0–999) are transformed to a single token. That straightforward change within the URLs illustration dropped the prices of the analysis section by greater than 50%!

There are most likely extra methods to scale back prices. I’m nonetheless taking part in round with this, studying how you can optimize it higher 💪.