Questions and Answers
Hello Don Jones
Thank you for writing this book
I want to ask about communication.
As a data scientist we understand that our models and predictions will never be 100% bcoz we always deal with sample of data and not whole data (aka population in statistics)
When speaking to stakeholders or customer who understands only business (basically doesn’t understand these concepts) how will suggest to approach so that we could negotiate in terms of margin of error.
Say we ask for 20% margin of error but stakeholders or customers asks for 5% and now we want to bargain on this as we know 5% will be very difficult due to constraints present in that situation.
Teach them :). “I can do that, but here’s what would have to be true.” Or, “with the quality of data we have, that just isn’t possible.” Don’t feel obligated to say yes if you can’t truly says yes - but remember that negotiating is a two way street. If you give something you should get something in return.
I also suggest making sure you’re taking the time to understand their context. That is, they might understand “only business,” but that’s a really important perspective. Make sure you don’t understand “only tech”—that is, make sure you’ve taken the time to understand the business as well, so that you can help translate for them.
As a business leader, I’ll use your example to perhaps point out the difference. For me, someone coming to me with a 20% margin of error is basically coming to me with nothing. Emotionally, I feel I could throw darts at a board and get 20%; 20% margin of error is a 20% that the data is completely wrong and I’ll make entirely the wrong decision.
That’s where it can help for you to better understand the business side. “Sure, I can only get you +/- 20% on this data… but help me understand what your’e trying to prove or disprove. Maybe we can add some other data sets, and between their intersections, we can narrow this down for you.” They’re asking for 5% because they’d like some confidence in the data; when you offer 20%, you’re basically telling them to not be confident in the data. That doesn’t help drive a business decision - so what would help drive a more valid business decision? Focus less on the data and more on the business outcomes being sought, and then look to how you can help support (or not) various aspects of that business decision. And if you just can’t… it’s fine to say so. “We don’t have the data to support this with any level of confidence one way or another. You’re going to have to go with your gut.” Sometimes, business leaders just need to rely on their experience instead of data they don’t have.
Thank you Don Jones
I totally agree to this. Thank you for sharing this knowledge.
I can totally relate myself with context part you explained, as that’s what I have been doing but didn’t know how to apply in tech world. As it’s totally new field for me.
About the example discussed - 20%. If I understood correctly:
- Keep a range in mind rather than a point.
- Don’t disclose your margin before they do. Ask about how are they confident about the data.
- If you n team with little exploration that margin cannot be met. Negotiate by asking more data then in terms of margin ?
- Focus more on understanding stakeholder thoughts i.e. in terms of business, rather than restricting yourself to only data.
- Finally if we can’t figure out a way to achieve the results with the margins specified. Let them know.
Please correct me if I am wrong.
Main motive to learn about this part was to be able to communicate with people with only business understanding and get information from them in terms of tech requirements.
So, thank you for clarifying this 🙂🙏
Yeah, that’s pretty much it. Understand how the data is being used. 20% margin might be fine if the goal is to decide where to aim a small marketing campaign; it might not be if the goal is to decide where to invest half the company’s resources for the next year. If you don’t have enough data, either get more or just accept that you won’t be able to make a fully data-driven decision. It happens. Some things just aren’t always knowable!
Hi Don, what would you suggest as the essential things you should do/connections you should make when joining a new company to set yourself up for success?
I think reaching to peers out of your org silo. Ask if you can set up a 30min meeting to learn what they do, and what their teams do. It’s a great ice breaker and helps you understand the shape and function of the org more quickly. I love it when new hires ask me to have a call with them!
Thats a great open culture in the Org. Kudos.
How do they learn/come to know what you or anyone else has to offer? How does one get an idea of what areas/topics others are doing? Usually someone senior guides them towards experts… Search via Confluence, perhaps :-)
One recent new hire of ours just looked at the org chart (we use Workday, which constructs an org chart). She reached out to everyone at her level, and a lot of people a step above, and asked for 30 minutes. I’m sure her direct leader offered some advice (I do that for new hires, for example) as well. But it’s in the talking to me and others where they learn what’s everyone does, how the org fits together, and so on.
Hi Don Jones, I’m wondering if you have advice for minoritized people in tech, and how should they navigate they careers?
Would you give them the same advice as the majority or should they tackle their careers differently?
I know that I network differently, but I’m wondering if you have more specific advice from your experience and “perch”
Thanks so much for joining us for a Q&A this week!
I’m a white guy who just turned 50, so although I’m not heteronormative, most people don’t know at first glance that I’m not part of the usual club. That makes it harder for me to offer meaningful advice, because I don’t have experience to speak from. I can offer what I’ve seen others do, though.
First, I clearly see equality still as a major issue, even in companies that truthfully want to do better. Unconscious bias is very real, and people can’t change what they haven’t yet actively acknowledged. I think just knowing that kind of preps you for what you’re up against.
I think the advice in the book is universal - but for minoritized people, it’s still not enough. I’ve seen women, for example, who’ve managed to walk the fine line between reminding people of their accomplishments and bragging, and they’ve been super successful - it sucks that they have to continually point out what they’ve done, but it seems to be reality.
I’d suggest seeking out allies. I’ve tried to introduce myself to new hires and offer mentoring, advice, championship, whatever I can, so they know they’re welcome, they belong, and I’ll try to have their back if they need it. Seek that out - outright ask for it, if you can.
Such a huge topic, and as an industry one where we seem to perpetually do so poorly. I wish I could be more helpful.
Thanks, I appreciate the honesty and that no one has all the answers! Alleys have definitely been a huge part of my career so far. Very much looking forward to reading the other answers and your book 📚
Hi Don Jones, lets say you are in a meeting with your boss and you are in a disagreement on a certain idea related to work, lets say some architecture, how to make someone understand the point in a manner which is not implied rude, some people speaks in a very direct manner which makes people look impolite, although they are not trying to be but they just want to be clear about things.
I’d say you do have to learn to read the room.
I wrote fiction on the side, and when I tell a story to a Young Adult audience, it’s very different from writing for an older audience - even if it’s the same basic story. So you do need to understand your audience. If you need to put them at ease, step slowly into a conversation, etc.
Ask questions. Disagreements often result from a lack of shared context - ask questions that help you understand their context. What about their decision is driving it for them? Are there factors you haven’t considered? Rather than making statements, act as if you don’t have all the information and try to get them to explain.
And at the end of the day, it isn’t always your decision. Present your facts, explain how they fit the criteria as you understand it, and list your assumptions. At the end of it, accept the decision and move on. It isn’t personal (or shouldn’t be.)
I often find that it’s easier to come across as someone who doesn’t have all the facts and wants to learn them; it avoids putting the other person in an “inferior” position, and when you’re genuinely trying to learn, they’d have to be jerks to not help out, right? (And if they are, maybe you’re not at the right company for you.) I don’t mean asking challenging questions like, “can you explain why you feel that way.” I mean asking questions like, “So, I may not fully understand all the business criteria. Can you help me see how your way is hitting the right criteria? I’d like to be able to offer better recommendations in the future, and having your context on this will help me be better at that.”
Hi Don Jones,
How would you approach role switch in your career? Example: Developer to Data Scientist role transition or ML Engineer to AI Researcher. How would you prepare for the job interviews? How would you utilise your network for this ? How will you validate your skills with respect to the industry in this new role which you want to work in?
That’s going to differ so much based on the exact role. I’d be very careful about reading the job description, try to speak to existing employees (hello LinkedIn!), and gather as much info as possible.
Certifications can provide a certain level of validation, but it’s usually minimal. You simply can’t “validate your skills” in most cases. You need to have confidence, and ideally have worked on some public or community projects so you have a portfolio to point to.
How do you incorporate your side projects which involves cutting edge tech into your job? Example: You’re working at company A and company A only applies some simple rule based approaches. You have learnt deep learning and want to apply into your job at company A. How do you convince the stakeholders at the company to pursue your interests so that both of the parties are benefited ?
I’m not sure why company A would even want to, honestly. It’s not their role to provide your career with opportunities or to keep you interested. If you can articulate a business-focused, data-driven argument, I’d hope they would listen. But focus on business outcomes, investment cost, ROI, and business factors. But just because you’re interested isn’t a valid business factor. If you want to pursue something new, why does it have to be at work? What kind of community projects could you start or contribute to instead?
Thank you Don Jones. This makes sense
Hi Don Jones : I bill myself as a generalist, and have been warned this is a hurdle for my career development (at a big company) unless I want to be a manager. Do you have thoughts on the generalist/expert divide? Is it really as binary as specialist or “manager”?
It kinda is in all but small companies. There’s just a feeling that a generalist won’t be able to do as good of a job, and the fact that big companies like to know the whole shape of a person and how they fit into a specific role. It isn’t always morally right, but it’s how the world works. Consider working for smaller companies or startups, who tend to put more value on a few clever generalists in their ranks.
I’ll offer what’s hopefully a good example:
I’ve worked for companies big and small. One tech shop I worked in, for a big retailer, was maybe 12 people. Small! They were delighted to have me as a generalist. I did programming, ran an AS/400, ran the phone system - a lot. They didn’t need a lot of generalists, but having one or two was a great fit.
Another was huge - maybe 1k people. Big! They had very rigid job descriptions, and they needed people to stay in their lane so as to avoid disrupting others.
It’s very much about the company.
I also find that generalists have a lot of difficulty describing their business value. “I can do it all!” Is not terribly convincing or believable; it sounds like bragging. How could we put you on a project for six months and then move you? It’s why startups are often more open to generalists—especially if you’re good at standing up new teams and then making them autonomous and then stepping away. That takes generalists, but few articulate it in the sense of “what kinds of business outcomes does this create.” And even that’s more a management example, I guess.
Actually GOOD generalists aren’t that common, so companies don’t create positions for them, and so companies don’t want to hire them. I’m actually a VERY effective generalist… but I can use that to specialize into whatever field I need at the time. And that’s wound me up in a VP seat :).
Thank you for the thorough and thoughtful answer. And the conclusion that, yeah, it tends to get you in a manager type role. :)
Second question :). Don Jones
Question: what advice do you have for people who want to get into tech, have a math/science background, and don’t know where to start/what in the tech space interests them?
I’ve suggested podcasts, informational interviews, etc, but there are just so many roles, company sizes and types, etc. And maybe the person is burnt out and nothing really entices them. How do they find something sufficiently ok? :)
Background: I left academia for industry 4-5 years ago and have jumped into lots of short term roles to try them out and find my direction, which I’m settling into. I now have other former colleagues coming to ask me advice, and I kind of wish my transition had been more smooth and would like to help them achieve that (see, e.g. CJs interview last week with Alexey on transitioning).
Yeah, it’s hard for sure. I think talking to people is something that can help - heck, start a podcast and start interviewing people from different roles, asking them to articulate what it is about their jobs they do, they like, and they do not like. That way, both you and your listeners can start to understand those questions. But it does take a lot of exploring.
But I mean, you’re kind of asking the same question that a high school senior might be asking. And you took some of the same approach that a high school senior took. When you’re entering something new, you just don’t know any of those answers, so you have to experiment a little. That’s why I always encourage younger people to try a lot of different jobs, before they settle on one that will be their career, or even before going to college, and majoring in some thing. Despite the fact that you might be 30, or 40, or older, you’re at the same point in the career as they are.
I’m not sure if there’s a way to shortcut that apart from a lot of rigorous research and talking to people.
Even thinking about a roll like software developer… That is so different across different companies, different technology stocks, different industries, different languages, different everything. Unless you have spent some time and started to experiment, it can be really hard to figure out.
Now all that said, I will offer this: community projects can be one of the best ways to experiment without committing. If you think you might be interested in being a software developer, find a community project that speaks to you and try contributing to it. It’s just like having a job, only it doesn’t pay. So it’s a way to get a feel for the job, without having to go through the commitment of interviewing, applying, getting the job, and then finding out you were wrong. All across the entire technology industry, you will find community projects that are desperately in need of contributors, many of whom are doing really amazing work for the world. It’s a great way to get your feet wet, and a fantastic way to build a portfolio that a potential future employer could look at to learn more about your technical skills.
I should also point out that I run https://ampere.club, which is a free follow-on resource to the book. I write weekly articles (all free) and there’s an audio (“podcast”) version for most articles written this year and in the future. You’re welcome to drop by and see what you think. There’s an e-mail newsletter option, which sends each new article to your inbox, if you prefer.
Hey, Don Jones. I greatly appreciate you doing this.
What general advice do you have for an older techie with a recent gap? I feel that my career is stalled.
I mean, what’s “stalled?”
Careers aren’t sharks that have to keep moving until they die. They’re like… benevolent kidnappers. They take you someplace and keep you there. So if your career is giving you the money and time you want, you’re there. You just need to stay there.
Apart from that, what problems do you solve in a business? Are those problems relevant today? Then no problem. Are they not? Well, you’ve got some learning to do - you need to solve a problem that companies need solved.
This isn’t just skills, like knowing the latest version of SQL server or whatever - it’s about being a solution that a company will pay money for. Think of yourself as a vendor. Who is your market? What do you do for that market?
So I ask that question of people a lot, and I’ll get replies like, “I’m a Microsoft Exchange admin.” That’s a job title, not a solution. What problem do Exchange admins solve? They keep servers running. They keep messaging systems running. Okay - that’s a solution, of sorts. Do companies need that solution?
Increasingly… no. They don’t. So that’s a problem. That means people who weren’t keeping their eye on the ball now solve a niche problem. So they’re gonna have to learn to solve a new problem—and find a way to convince potential employers of their ability.
They best way? Get involved in community. You get to learn new skills, solve common problems, and you create a public record of your expertise. A portfolio, something way way way too few technologists take the time to do. But working on community and open-source projects is a fantastic way to demonstrate the problems you solve. Yeah, you may need to learn new stuff… but I mean, that’s tech. People who don’t like to learn new stuff don’t work in tech, they work in lumber (haven’t had a new kind of tree for a minute). You’re in tech, so you can learn to solve problems, demonstrate that ability in public view, and leverage that into whatever you need your career to be doing for you.
Don Jones can I quote your comparison of careers and (not) sharks? Brilliant analogy. Obviously would attribute you (was thinking a tiny LinkedIn post referring to the thread/your book, with said quote).
Absolutely. Please do!
Don Jones I read this 🔼 response with interest. I’ve had the experience of being with a benevolent kidnapper. I had a good job with a name brand company, great pay and happiness. However, I never realized when I became a dinosaur.
“You just need to stay there” described above seems like a lot of work. Creating a good portfolio project and online presence takes a lot of time and effort, which, in addition to full time work takes time away from family, kids, friends, volunteering, chores, and just relaxation.
Tech, in a lot of ways, feels like a hamster wheel - Keep running faster and faster or fall off. This, I speak as I spend 3-4 hours every week upskilling. Why? The pay is good - but I do not want to repeat a scenario where my skills are outdated again.
Any advise for a jaded person here?
I mean, if you’re not into constant change… then tech isn’t a great field. It’s why old guys like me often move into management. Those skills last longer. I also hit a wall with being a tech practitioner, so I turned those years of experience into something different.
Yeah, keeping up with tech is a lotta work. For sure. And it never does stop. But that’s a lot of fields, right? Law. Medicine. Really any field that’s dynamic - and they mostly do pay well in exchange.
Moving into management also means taking responsibility for outcomes. Maybe it’s a progression, but do you have any suggestions for thinking that way. I understand one can never be fully prepared.
Yeah… I mean, in good companies, entire teams are responsible for outcomes, right? Leaders are there to clear the way for their teams to do the actual work, to set a vision and facilitate creating a strategy. Good leaders recognize that they produce no work output and that they’re there to partner with their teams. If you’re an individual contributor and you don’t feel “responsible for outcomes” already, then you’re not being led very well. For me, the best leaders ar ones who can communicate what success looks like and help every team member see themselves in that success, and show everyone, “hey, if we do these things, success will happen.” It’s not actually hard, but it’s a distinct skill set.
Hi Don Jones thanks for the Q&A! I’ve found that making time to intentionally improve soft skills seems to be more difficult than for more technical skills, and I think it’s because “Improve my presentation skills” has a less definite outcome than “Learn X tool”. Do you have suggestions for how to frame the goal of learning these skills more concretely so that the allotted time for it doesn’t fall victim to more immediate and tangible demands.
One of the things I’ll be doing on https://ampere.club in 2022 is dedicating each month to a “skill theme,” and providing some actionable outcomes and tasks for each every week. So that can be a way to help.
But set yourself milestones. “Give a lunchtime presentation where I’m rated at least 3/5 by the audience,” for example—actionable, measurable, and contributes to improving a skill. I’d actually say “Learn X tool” isn’t all that great; “Learn to do X with Y tool” is far better, and if you think that way, then soft skills are just tools used to do actual things. Measure the things, not the tool.
Thank you, great suggestion!
Hi Don Jones thanks for doing this Q&A. My question is.
How will you evaluate a tech idea when you are starting a company. Say you are starting a quantum computing tech related company which might niche for the market, or might not give you the ROI right away. But you are really passionate about the tech.
”Passion” isn’t really a great reason to start a company. I know people say that all the time, but the best question is, “what problem am I solving, and what’s the best set of tools to solve that problem?” Then you can look at how many people have that problem, how much they might save by paying you to solve it for them, etc. Otherwise it’s a hobby, not a business. And hobbies are fine! But they’re hobbies.
Hi Don Jones, thanks for doing the QnA session. We all know how important communication and being able to explain our work in simple words to layman is in the tech field. How would you suggest to enhance these skills with time?
First, practice, practice, practice. For example, start writing articles—even if you don’t publish them!—that explain complex technical concepts to a lay audience. Actually, that’d make a great blog.
Remember: teaching is nothing more than repackaging information for a specific audience. Teaching doesn’t create new information; it places existing information into someone else’s context.
So next, you have to learn more about other people’s context. Get to know your audience in general. What’s their background? What shared experiences do you have? What analogies tend to ring true with them? If you want to teach someone, the burden is on you to do the work to understand where they’re coming from, and to construct explanations that work for them.
In business, that also means taking on the work to better understand the business. Growing your business acumen, understanding business drivers and motivations, all of that is also very important. It’s a big part of what I do at https://ampere.club, in fact—focus on business acumen-building.
Also, can you please suggest some ways for an individual to comfortably ask questions from colleagues or mentors….we tend to be scared that people will judge us based on our questions.
That’s a complex question, and it comes down to “fear of failure.” None of us wants to look stupid, and we believe that asking a question is admitting we don’t know something, which makes us look stupid.
Which is stupid. Nobody expects us to know everything.
If you don’t know something, learn it. Now… keep in mind that you colleagues also have jobs to do, and they don’t necessarily have time to answer a million questions. So make the effort and investment to self-educate as much as you can. Once you reach the end of that effort, you can ask informed, intelligent questions that show you’ve already made the effort—honestly, I think most of your colleagues will appreciate that. And if you work in an organization that doesn’t, maybe you should ask yourself why you work there.
Hello Don, could you provide any tips to find out when one’s career is stalling and is time to move up the next step? I know it’s a bit vague but it’s hard to know sometimes if one is falling behind or progressing within one’s potential. Thanks 🙂
Well, I’ll refer you to the first chapter of the book. You need to decide what kind of life you want to live, and then determine what kind of career is needed to support that life. There was a previous similar question here about a “stalled” career.
From there, as I wrote at https://ampere.club/what-problem-do-you-solve/, it’s then about “what problem do you solve?” Your career is about you, but the jobs you hold are about the employers who pay for them. If you need a different job, then you need to look at the problems employers need solved, and then skill yourself to solve those problems. Think of yourself as a vendor: what do you bring to the table that employers are looking for?
A way to discover that is to browse job listings and look for common themes. Those are clearly areas of opportunity.
But your career is never “stalling.” Either its taking you toward your success, or it isn’t. If it isn’t, you adjust your career plan. If it is, you continue executing your career plan. If you don’t know, or you don’t have a career plan, you read my book 🙂 And get a career plan put together.
thnaks, I’ll definitely buy your book (unless I win this week) and spell out a plan
Hi Don Jones I’d be happy to hear your thoughts on
- lateral career moves in general. But to get it started, Is there a timing element to consider? or maybe specific moves you saw that worked well or didn’t? was it about domain combination?
- Moving from management to IC (and maybe doing it several times back and forth).. I am enjoying taking breaks from management, and refreshing my technical skills despite my love of management. Is this unique, or is it a career path pattern?
So, on laterals… not really. It’s no different than any other job move.
On the other… I mean, it depends on what you’re trying to get your career to DO for you, right? Going back and forth like that isn’t a sign of a focused career that has an outcome in mind, it’s a sign of someone just kind of following their nose. Imagine getting into your car, turning on the GPS, not entering a destination, and then hopping on and off the highway. Where are you going to wind up? Sure, you might great sights along the way, but eventually you run out of gas… where?
Your career isn’t thee to serve your interests of the moment. It’s there to create a life for you (work to live, not live to work). It can’t do that if there’s no plan.
People kinda act like “there are technical skills, and then you move into management.” Management is a set of skills. It’s a distinct career path. It’s HARD, when it’s done right. I’m not sure you can be the best manager your people deserve if it’s… a hobby that you set aside when you don’t want to do it anymore. If you’re in management and you still want to keep your hand in tech, find a deserving community project that needs you. Find a “Kids Who Code” charity and offer to be an instructor. There are lots of ways to “stay in tech” without making it a fractured career.
Hi Don Jones thank you for all your answers so far. I was wondering if you have some advice on how to help improving a team from within. For example, motivating the team to commit to a certain set of minimum standards concerning code standard, PR, etc.. Also how would you motivate your colleagues to reach certain KPIs.
What motivates YOU?
Watch this: https://www.youtube.com/watch?v=fW8amMCVAJQ
That’s “leadership from within.”
Your colleagues should WANT to reach those KPIs, because it’s what they get PAID to do. Where you can help is showing them that it’s achievable. Showing them that it matters. Showing them that it’s WORTH doing. That’s not going to work for everyone, and those people should consider leaving, but it will hopefully work for a lot of people. Show them WHY these things matter. be passionate about it—let them hear the excitement and conviction in your voice. Don’t lecture… show. Show the upsides of code standards… don’t just talk about it. That kind of thing.
How to best collaborate with young highly motivated colleagues that are still missing a lot of the technical skills and with more senior colleagues that might have lost a bit of their drive?
Remind the senior people what it was like to be young. Remind them how many people took time out of their day to help THEM. They owe this to those people—paying it forward. We’re all just stupid monkeys on this planet; we need to help each other get along. Acknowledge the younger folks’ excitement and spirit—don’t beat it up. Life will beat them up plenty, you don’t need to help. Heck, maybe they can help raise the bar for everyone, in time—if their elders can set the right tone and give them the right support.
For the younger folks, let them know how interested you are in helping them. Let them know it’s okay to not know everything, and that it’s far more important to be a #DailyLearner. Let them know their questions are welcome. Let them know their perspectives are welcome—and that they need to seek context to understand why things are the way they are, but beyond that… they’re welcome to think about better ways that meet the same goals.
thank you Don Jones I will definitely try to act more in this way
Hi Don Jones! Are you on LinkedIn? I was hoping to follow you for awesome content and your name is a bit too common for me to quickly find you (figured I’m not the only one interested, so makes sense to post instead of dm).
Ok, one more question. Don Jones, you said in your answer above that your generalist nature did in the end land you in a VP role. Could you say something about that transition ? Are your generalist skills still your biggest asset, did they nicely power up your acquiring of management-related skills for such a leadership position?
So, I ran my own company for about 14 years. Being a generalist is 100% key in that scenario—you are the Finance team, the Sales team, the Production team, everything. I still think those generalist skills, as a VP, are my biggest asset—I can dive into a variety of topics, understand them quickly, and contribute meaningfully. It lets me interact and engage with a variety of teams in a variety of ways and really Get Stuff Done. Management skills are one of the things I’m a generalist at, and I’ve become more and more competent at those as I’ve worked in management—but I can also write a PowerShell script, when I need to ;).
You’re all welcome to follow/connect on LinkedIn (https://LinkedIn.com/in/concentrateddon) and Twitter (https://twitter.com/concentrateddon), and I’m happy to take future questions (I watch Twitter the most, but please post publicly versus a DM so everyone can play along). And as I’ve mentioned, you can also find me at https://ampere.club, and there’s an “Ask” page you’re welcome to use anytime. For 2022, Ampere Club is going to have monthly career-skill themes, with small weekly activities designed to help you level-up in continuous, minor ways (like Agile for your career!). It’s all free, of course.
thank you, really cool stuff!
I want to thank everyone for the fantastic questions! I’m going to genericize some of these and post longer answers, in the form of articles, at https://ampere.club, because I think your perspectives and questions reflect those of a much wider audience. Hopefully, my extended answers can even provide a little bit of extra value, and hopefully you’ll continue to pop by and ask questions!
Thanks Alexey Grigorev for yet another interesting week.
Question I have regarding the book:
“What according to you is a major drawback while using fastai as compared to TF or PT (apart from the fine grained control we have from these 2 frameworks)?”
Again kudos Mark Ryan for such a detailed book, the TOC looks quite comprehensive 👍
Hi Raghav Bali - fastai is a great platform, particularly for beginners. That being said, I see two noteworthy drawbacks: (1) compared to TF/Keras or vanilla PyTorch, there are not as many examples of fastai being deployed in production, (2) regressions. Compared to Keras, it is easier to hit regressions with the fastai platform (that is, something that used to work that stops working after a platform update), and you need to be prepared to work around regressions if you come across them in fastai. An example of a regression in fastai is the code to set the random seed, which stopped working as expected after an update. In addition to these drawbacks, the documentation for fastai is not as comprehensive as Keras documentation, but the fastai course forum https://forums.fast.ai/ helps to make up for that.
Awesome… Thanks for the detailed answer