Congrats Geoff! Your advice and the empathy associated with it has been very helpful as we evolved out of YC. Glad to know the future batches will continue to be in great hands :)
BioPhysics was all the rage in my university few years ago. I haven't heard much after, but genetics and big data have been at the forefront for a while now. There have been a bunch of recent YC companies in this intersection as well.
> I run the team that builds software for Google's quantum computers
Is there any significant difference to payscale and hierarchical autonomy in these non AI research teams? I am assuming distributed privacy preserving ML team (is this team close to federated learning one?) also falls under more non AI research teams right?
I have a bachelors and Masters (2013, with Non Linear Dynamics as MS thesis) in Physics. Have always loved the field and the mathematical rigor involved. However, chose a strategic consultant job as the pay was very lucrative. Got disillusioned in a year and pursued a masters in CS/AI and worked after for an year in ML. Started a company and got acquihired into an engineering managerial role after.
Looking back perhaps one major regret I have is how far and fast I am moving away from fundamental sciences. I don’t think I can leave the bay area or coding anytime soon, but nonetheless, I have started to look out for ways to stay involved with the world of physics in as many ways possible
(OP here)
Quite amazed to see the number of physicists here and it is very heartening to see so many of you doing so well in such a wide variety of fields. Especially since during my college days physics was considered to be a slightly dangerous choice from a job/career growth perspective
I often question if putting so much blame just on her while treating her investors with a lens of victoimhood is a good idea? For all I know, Theranos was a vicious death spiral which grew so big purely because there were a ton of investors writing big checks to Holmes either without doing basic due diligence or being knowingly complicit in the lackings of the company. Their checks were the reason Holmes got empowered so much in the first place and dared to think that she is perhaps on the right path and could continue pulling off a fraud.
As an ex-founder who was mad about side projects, I can tell you in hindsight, if you start a side-project, definitely make sure you enjoy what you are doing. The joy is in the process and not the destination. Otherwise, you will just burn your way through and destabilize your relationships unnecessarily as well.
Can you elaborate on why you think this might be due to screen exposure? Just curious since I often have a vision pain too, but was not sure if that is indeed because of screentime. I tend to take time off from screen at regular intervals and maintain healthy lighting around etc.
This looks absolutely amazing! I would be very curious to know how you went about conceptualizing the project and the AI beneath. Do you have a blogpost on it or planning to write one?
At 800K, I am extremely curious how your work-life balance is? I am assuming after an age and salary number, a lot of your time goes into managerial tasks and at the rate you are compounding on salaries, I would bet it will be a hectic WLB. Further out of curiosity, at what age do you plan to retire?
I can turn lights off without getting off my bed, and damn, has it made my life crazy simple. I was in India for few months last year and I realized a large part of me getting up early was to just switch on the geyser, open the door for the cook and turn the lights on. I would get back to bed after and check up fb till the water got hot. Just the fact that Echo can do these things without making me get off the bed will be a crazy value addition to my current decently privileged life.
filter and formatter are still set at the time of workflow definitions and are not real time. When I say real time, I mean that the user in our case can see the exact event (like say the content of the email, ticket or crm update etc), and then decide to take an action. And when he/she is passing event data to that action, they can update it based on what they see and not what they expect as is the case with other automation platforms. In effect we are more interactive workflows than automated workflows if that helps.
yup. makes sense. Jira does provide those. The main element missing in native slack apps is the ability to send data across different integrations. So, some cases where a cross channel support that we provide might be handy is if you want to create jira ticket from email, update sheets from jira notification, reassign tickets and update calendar based on email etc, but not sure how often you will need something like this. but if you do, let us know and I can reach out to you directly.
Thats a very good question. There were 2 main reasons for that. There are a lot of chatbot solutions to enable human initiated workflows while there were hardly any solution which enabled human in between an event driven workflow. But more importantly, since we came from a sales focused market and faced the problem of supporting human element within an event driven workflow ourselves, it was easier for us to get the relevant customer feedback.
Also, chatops right now is mostly human initiated and is focused on tech. We wanted to tackle enterprise integration market where the need seemed more immediate. A lot of the people we talked to hated logging into their CRMs and updating them at regular intervals can be a very tedious task. Hence we decided to prioritize event driven workflows with human in between, while keeping the architecture open to support human initiated workflows triggered from slack as well.
Hi sure. Fair question indeed. We were a sales call scheduling service till september 2018 and had built a strong email categorization and reply layer on top of collaborative inbox to manage multiple client inboxes, schedule calls and update CRMs. By late last year we decided to push to a pure SAAS play and eliminate the human-in-the-loop segment of the company. You can think of this as a soft product pivot to achieve scalability and reduce operational overhead.
That said, since we allow AI enhancements on the event data (email categorization to be precise for now) on top of the zapier model and also allow the human decision making element in between, we could not just plug and play any IFTTT architecture, which definitely meant some more engineering overhead.
hey, thats a very good question and one that we get very often.
Zapier is a complete automation suite where in an event will automatically trigger a bunch of actions without the ability to modify the event data or decide the action based on the event.
We enable all these usecases where a user will have to see the event and then chose which set of actions to trigger. the user can also send a modified version of the event data to the actions, while in Zapier, the event data consumed by the action layer cannot be updated in real time.
Haha, Thanks for pointing it out. Will change it soon. One of the biggest reasons we chose slack as the notification consumption platform is to enable collaboration(and therefore more visibility across teams) as we expand to more usecases outside of sales. We therefore built a robust email <> Slack integration as well.
Also, one of our main USP is the ability to take action on any notification. Slack provides a strong api layer to pull this off, while its extremely tough to do the same as a gmail app. Chrome extension is still a viable option, but for now, we chose to go ahead with Slack due to the collaboration element.
Regarding the engineering manager usecase, although we don't have deep relevant integrations yet, we are considering adding support for Jira, Trello and Sheets as well, but would love to know if any of that will be as helpful in your day to day work. We are not venturing into tech alerts yet, as there are good comprehensive solutions already in place but would love to know of any gaps that might exist where a configurable action layer would be helpful.
These are all frameworks and while none of them have any signular advantage over other especially in the problem statement you are looking for, you should ideally be able to figure out what works best for you based on the classification sensitivity and training data you are working with. The problem in itself can be quite simple to extremely complex based on the above 2 factors. Spacy's pre-processing tools are quite easy to use and that combined with tool like talon should help you clean up the email correctly. Thereafter, if your email text is pretty much to the point, then any intent classification tool will work, however, if the email text is long and intents are spread across, then you will need a hierarchical layer to understand the intent hierarchies as well as an attention layer to understand which intents to focus and not lose track of in an email. At that time, you are quite far from using a generic plug and play framework and will need to exactly and quite thoroughly understand the deep learning models you are working with as well as the dataset you have and the classification you are trying to build.
Working in software development in India after working in bay area for 5 years. I cant speak for China, but the work Indian companies get is singificantly less critical than what their US teams are working on. As a result, there is much less emphasis for best practices and good coding culture here in comparison to US. However, considering how many of my Chinese colleagues have returned back to China to code with local startups there, I would like to bet that the Chinese coding culture is far more closer to bay area than to India.
Kudos to YC for taking this approach. A lot of founders want to quit status-quo and pursue startup full time but are not able to due to a number of reasons including college, immigration, teams being risk averse etc. With this, they get the validation that they have a good investor backing them and also get the time to test out their PM fit before getting into the batch, giving them enough time to learn and iterate on their mistakes.
I always wondered how the stock exercising process feels like for an early employee who has left company few years before the IPO. Is there any feeling of lost opportunities amidst all the happiness of the exit?