Few days ago I tweeted things I should not have. It was bad, I regret and apologize.
This sort of stuff undermines the effort of colleagues, and my own, to articulate the important role various disciplines play in taking ML forward and to create a welcoming and healthy community.
This sort of stuff undermines the effort of colleagues, and my own, to articulate the important role various disciplines play in taking ML forward and to create a welcoming and healthy community.
I meant to call out a person for a pattern of behaviour. I should not have done that, I did not give enough context, I worded things poorly, the rest is history.
Most painfully, I turned a personal thing into an us-vs-them message which I do not otherwise endorse.
Most painfully, I turned a personal thing into an us-vs-them message which I do not otherwise endorse.
Let me try to constructively address the situation and repair some damage with a positive message.
Research, research engineering, software engineering are all important skillsets and contributions and deserve respect and recognition. I agree with this, believe this, I say this often, and I am aware that my tweet communicated the exact opposite 🤦.
One can drive ML forward in many ways. You can possess great software engineering skills, be good at theory, or be a first-principles thinker. You can be good at empirical research or beating SOTA or communicating ML. Or, you can use leadership or power to further ML.
These are not mutually exclusive. Some people are good at all of the above, while many successful people tend to be outstanding in a subset of areas. This is hardly surprising and is, in fact, a great textbook example of explaining away.
Being good at one discipline doesn’t disqualify your from being or becoming outstanding at another discipline. My tweet implied that I believe the opposite, and I don’t. I don’t think anyone should believe that.
I believe if you contribute great things as a software engineer or as a researcher we should recognize that. And we should create every opportunity for you to become a better researcher or to get into research. My tweets sounded territorial, exclusive and stupid.
As people pointed out my tweets contradicted the very thing I am trying to achieve with my blog: allowing people to sharpen their research intuition, broaden their knowledge, brush up their technical skills and exercise critical thinking.
Research is ultimately about seeking knowledge. Ideally done with rigour and with openness to criticism and collaboration. It is not about maths, equations or theorems. I failed to convey this message which I wrote with more clarity earlier: 
I, for one, am a mediocre software engineer. I am also not particularly good at empirical research and running experiments. I am known for first-principles thinking and for communicating insights on my blog. Few readers could name a paper or research contribution of mine.
Let me also emphasize that ML research/engineering community is generally lovely and not hostile. It’s perhaps a few people, evidently now including myself, with wide reach who create a lot of noise and the unhelpful impression of a hostile and unwelcoming environment.
I hope these uncalled-for fights don’t deter you from learning more, wanting to improve, and engaging in constructive discussion. I learned a lot from this episode, the hard way. I’ll think more about what I write and resist the urge to get involved in every debate.
I recognized how stupid my tweet was right away and contemplated deleting it but made the call that deleting a tweet that others started responding to is bad practice. In hindsight, I think this was a bad judgment, deleting it would have prevented spreading an unhelpful message.
