Vector DBs for Data Protection
I was reading up some information on vector embeddings recently and started thinking about different use cases of it. One interesting use case was for Data Detection and Response. The thought exercise I had was that if we are able to track data access patters or data lineage and represent them as embeddings, then the embeddings of any anomalous / nefarious access pattern would have very little similarity with other embeddings.
Using External Functions with OpenAI
An easy way to boost the capabilities of AI models is to provide it with external information that it may not have. e.g. maybe your company has some proprietary data that would never be included while training the models. But if you could provide the model with this info during inferencing, it would be able to provide more accurate results. One way of doing that with OpenAI is via function calls.
Having Fun with OpenAI
AI is all the rage these days. New capabilities, that seemed impossible just a few months ago, show up almost weekly. So I decided to take it for a spin and see what different tasks I could code up quickly. I chose OpenAI, but you should be able to replicate these using any of the other models as well. Printing the text of a html page Book Recommendations Converting csv to json using json_schema Printing the text of a html page The first quick task I tried was to ask AI to print out the text of a html page i.
Measuring Patterns To Boost Productivity
Today the most common way of getting feedback is via peer reviews. Some companies do it annually, some bi-annually. This helps employers to reward the most promising employee through promotions, salary raises etc. One disadvantage of the above format is that the feedback is essentially an evaluation by peers of past performance. So it’s already too late to possibly course correct. The next time around the peer review could be done by someone else, with different internal yardsticks of evaluation.
Will AI take my job ?
Whenever a new technology becomes commonplace, there will be disruptions. Technology and their associated disruptions have occurred throughout history. e.g. The monks / clergy were primarily responsible for making books (by hand). This restricted the flow of knowledge. Then the printing press came along. In the 1920s, telephone companies employed thousands of folks as operators. Today algorithms have largely taken over. Essentially any manual / repetitive task was automated out of existence.
Rates of Change
I don’t think we understand rates of change very well. The reason AI is all the rage now is that it offers an opportunity for a gigantic leap in the rate of change of innovation. No one can say what it will usher in. But it seems like an ideal vehicle to bring about this big change. Big rates of change is what brings about disruptions and opportunities. The big players can’t possibly address them all.
Getting Through Your Day
Some days can be exhausting. For many, most days are exhausting. We just think that it’s the way things are and plough along. Same thing, day in and day out. That’s life ! On one end of the spectrum things seem like they are exhausting because we are bored of it. It no longer excites us. The monotony carries over into boredom which carries over to exhaustion. On the other end of the spectrum are the passionate ones.
Assistive Vs Transformative Technologies
Software is eating the world. Or rather, software has already eaten the world. There isn’t an industry that has not or will not be disrupted by tech. But all tech is not created equal. Some help in making things more efficient (assistive) while some absolutely disrupt entire ecosystems (transformative) The internet began as a DARPA project. Initially it connected only a few universities. Today it’s pretty much a fundamental right. Since it’s inception it has transformed everything.
How To Do 1:1 With Your Team ?
One on Ones (1:1) with your team members may be the single most important tool available to a manager. But it’s often not done properly and consequently its benefits are not realized. But first… NOTE: This post talks about 1:1 from an engineering manager’s perspective. But it should map to other domains as well. What’s a 1:1 ? A 1:1 is a recurring time slot deliberately set aside so that a manager and a team member can sync up and discuss successes, challenges, personal goals, product roadmap etc.
How To Prioritize Tasks ?
Shipping products is hard. What makes it hard is that typical products involve multiple teams and multiple dependencies. Navigating these challenges is non-trivial. There are technical challenges to overcome, but those are typically not the biggest blockers. (There are enough smart people in the room to take care of that, unless its some greenfield domain like AR/VR/Self-Driving etc), Then there are the procedural challenges; what should be done when and by whom.