m98789

I have an “out of the box” solution:

Raise your price!

When your contract is about to end, let them know the price has increased to level set with the market. Even at 50% higher, you are still under market, so you can comfortably charge higher. They don’t have any other options if they want the same level of effort.

Now you have two outcomes:

1. They agree, and pay more
2. They don’t agree to pay more

If (1), you win. You can distribute the work to another person on your team and continue to grow the business.

If (2), you still win because you can now confidently drop a client who has clearly been a productivity drag on you, limiting growth (at 20 stories and 4 posts a day there’s not much else one can do).

The interview process for DeepMind is a bit different than elsewhere in that there is a heavy emphasis on mathematics. You will be quizzed on your maths.

Yes. Even the authors of the AI frameworks like PyTorch aren’t usually writing the low level cuda code for NNs. They are wrapping the cuDNN library from NVIDIA which has highly optimized cuda code for NN operations.

Correct, even for training the models, all the Python code you see is really just a friendly interface over highly optimized C/cuda code.

There are no “loops” or matrix multiplication being done in Python. All the heavy lifting is done in lower level highly optimized code.

Python is just the glue.

Q* haven’t you heard?

That’s not how the enterprise consultancy business works.

It works initially through your personal relationships.

Then spreads initially through the relationships of your clients (referrals).

Then spreads through the relationships of your partners as you bring them aboard.

Relationships is where you need to focus, not YouTube videos. YouTube, blogs, etc is for product companies or if you are already massive scale, and is marketing. What you are doing is enterprise services, and what you need is sales activities, not marketing activities. So you need to get to your network, make calls, offer low cost initial services to get case studies to help on next ones, etc.

Trust is #1 for a partner. Attitude next. Competency third.

Phi-2

• Just say what it is in plain English, no ChatGPT fluff “an innovative platform that leverages AI…” is a major turnoff to devs.
• Communicate clearly the problem it solves, not just what it does.
• Provide examples of the key scenarios / use cases and the costs / benefits of using this tool compared to other methods or competitors
• If you want experienced ML engineers to put their time and effort helping you, either have a budget to pay them, or make all of this open source so they have at least a community impact.

That may have the opposite intended effect

Hf?