VentureBeat readers possible seen this week that our website seems to be totally different. On Thursday, we rolled out a big design change that features not only a new look but in addition a brand new model construction that higher displays how we take into consideration our audiences and our editorial mission.
VentureBeat stays the flagship model — dedicated to masking transformative expertise that issues to enterprise determination makers — and now, our longtime GamesBeat sub-brand has its personal homepage of kinds, and undoubtedly its personal look. And we’ve launched a brand new sub-brand. This one is for all of our AI content material, and it’s referred to as The Machine.
By creating two distinct manufacturers below the principle VentureBeat model, we’re leaning exhausting into what we’ve acknowledged internally for a very long time: We’re serving a couple of group of readers, and people communities don’t all the time overlap. There are readers who care about our AI and transformative tech protection, and there are others who ardently comply with GamesBeat. We wish to proceed to domesticate these communities by means of our written content material and occasions. So once we reorganized our website, we created devoted area for video games and AI protection, respectively, whereas leaving the homepage as the principle feed.
GamesBeat has lengthy been a standout sub-brand below VentureBeat, due to the management of managing editor Jason Wilson and the exhausting work of Dean Takahashi, Mike Minotti, and Jeff Grubb. Thus, giving it a devoted touchdown web page makes logical sense. We wish to give our AI protection the identical remedy, which is why we created The Machine.
We selected to take a protracted and winding path to deciding on The Machine because the title for our AI sub model. We may have simply put our heads collectively and picked one, however the place’s the enjoyable in that? If you’re going to provide you with a reputation for an AI-focused model, it’s best to use AI that can assist you do it. And that’s what we did.
First, we went by means of the mandatory workout routines to map out a model: We talked by means of model values, created an summary about its focus and targets, listed the applied sciences and verticals we wished to cowl, and so forth. Then, we people brainstormed some concepts for names. (None stood out as clear winners.)
Armed with this knowledge, we turned to Hugging Face’s free online NLP tools, which require no code — you simply put textual content into the field and let the system do its factor. Essentially, we ended up following these tips to generate title concepts.
There are just a few totally different approaches you may take. You can feed the system 20 names, let’s say, and “ask” it to generate a 21st. You can provide it tags and related phrases (like “machine learning,” “artificial intelligence,” “computer vision,” and so forth) and hope that it converts these into one thing that will be worthy of a reputation. You can enter an outline of what you need (like a paragraph about what the sub-brand is all about) and see if it comes up with one thing. And you may tweak various parameters, like mannequin measurement and “temperature,” to extract totally different outcomes.
This type of tinkering is a pleasant rabbit gap to tumble down. After incessantly fiddling each with the information we fed the system and the assorted adjustable parameters, we ended up with a protracted and hilarious checklist of AI-generated names to chew on.
Here are a few of our favourite horrible names that the instrument generated:
- The Artificial Intelligence publication
- The Artificers
- The New York Times
- The New New Thing
- BrainNuts Journal
- IBM, IBM, and IBM
- AIB ET N US P L AI
This is an effective lesson within the limitations of AI. The system had no concept what we wished it to do. It couldn’t, and didn’t, remedy our drawback like some type of title merchandising machine. AI isn’t inventive. We needed to generate a bunch of knowledge firstly, after which on the finish, we needed to sift by means of largely unhelpful output (we ended up with dozens and dozens of names) to search out inspiration.
But within the detritus, we discovered some nuggets of unintentional brilliance. Here are just a few NLP-generated names which can be really sort of good:
- The Singularity
- The Breakthrough
- We the Machine
It’s value noting that the system all however insisted on “AIBeat.” No matter what permutations we tried, “AIBeat” saved resurfacing. It was tempting to pluck that low-hanging fruit — it matched VentureBeat and GamesBeat, and there’s no confusion about what beat we’d be masking. But we people determined to be extra inventive with the title, so we moved away from that building.
We took a step again and used the lengthy checklist of NLP-generated names to assist us suppose up some contemporary concepts. For instance, “We the Machine” stood out to a few of us as notably punchy, nevertheless it wasn’t fairly proper for a publication title. (“Hello, I write for We the Machine” doesn’t precisely roll off the tongue.) But that impressed “The Machine,” which emerged because the winner from our shortlist.
“The Machine” has a number of layers. It’s a play on “machine learning,” and it’s a wink on the persistent worry of sentient robots. And it frames our AI staff as a formidable, well-oiled content material machine, punching effectively above our weight with a tiny roster of writers.
And so, I write for The Machine. Bookmark this web page and go to day-after-day for the newest AI information, evaluation, and options.