Mojo: First Impressions
> This is the power of launching early, you're building out in the open allowing your users to witness firsthand the team's proficiency, building trust and enthusiasm as you ship features and fix bugs, while giving you valuable feedback about what your core audience actually wants.
Its eye-roll inducing.
My two favourite phrases that were totally not written by a bad PR team:
> Each time a blog post or video is shared the Mojo team respond to it with enthusiasm, which engages the creator as they become more involved with the community.
> Syntax debates have already begun, which have surprisingly been met with a high level of maturity, once a decision is made the bike-shedding has ceased.
Thanks for the feedback, on reading back the blog post it did read a bit like a "bad PR team", I've rewritten the article hopefully it's better :)
It is a bit weird how there's no credited author of this - as far as I can tell the posts are written by https://github.com/mojodojodev - but they still represent opinions in terms of things like "My only criticism on the launch would be".
Adding an author name would definitely help out with credibility here.
For every new technology there are hundreds of previously unknown tech blogger/influencer/entrepreneurs trying to cache in.
Imo it would be more pleasent if their content had more depth. At least this one had more compated to the x/10 twitter threads about promt engineering or graph databases.
In every new city in the West one of the first settlers is always the hardware store.
Gotta sell those shovels and dynamite barrels while the gold source is flowing.
> while the gold source is flowing
Even if it never was...
When I finally got access to mojo, I was very unimpressed to say the least. For how much marketing and attention they’ve gotten already and the size of their team, the product is in very early stages. It’s not clear how their product is different at all from the many existing mature (and free and open source) solutions for writing high performance code in python (cython, numba, jax, triton, ect…). They don’t even have gpu support whatsoever yet, making the whole thing pretty pointless and difficult to even call a “launch”.
It might be unfair, but I will measure this by how Swift for Tensorflow was sold by all its participants.
What is it about Mojo that drives people to write weird articles like this? I don't remember seeing so many bizarre hyped up puff pieces when Carbon was announced.
> What is it about Mojo that drives people to write weird articles like this?
Having some familiarity with marketing and PR tactics, they're likely being paid to, and maybe even work directly for the people behind Mojo. Note that the site repo is "owned" by the anonymous @mojodojodev and was created 3 weeks ago — strong sock puppetry vibes.
I'm not affiliated in any way or paid by Modular, in fact I criticize the marketing in the last paragraph, I'm just excited about Mojo and so the article sounded a little too puffy, I've rewritten it.
Go back reading when Swift for Tensorflow was announced, open.ai opinion on the matter, and compare with Mojo.
For anyone wondering: Mojo is a new programming language for AI developers, that supposedly combines the usability of Python with the performance of C, unlocking unparalleled programmability of AI hardware and extensibility of AI models.
> unlocking unparalleled programmability of AI hardware and extensibility of AI models
In peasant terms what it means?
Because they've limited the dynamism of Python so you can't do as much crazy runtime nonsense, their (vaporware) tech stack is able to (theoretically) compile more of your program into GPU kernels.
And the second attempt at it, after Swift for Tensorflow adoption failure.
Even though the front page says "Unofficial and not affiliated with Modular or Mojo", this reads very much like an advertisement, especially since the author is anonymous and is invested in the language so much that they started a "mojo dojo" site for a brand new language.