randyburden

I just tried it out and it is a great improvement. I upgraded our product to use the new turn taking model a few moments ago and also noticed there were new voice models to select from, so I upgraded us from gpt-oss to gpt-5.6-terra, which I'm hoping should help with our more complex instruction following training scenarios. Great work!

I learned to program on a TI-83 and later bought a TI-84+ with the cable that allowed me to transfer my apps and games between my device and other students devices. I have fond memories of hand typing into a TI-83 BASIC for hours using code I found online at the local library - games like Drug Wars and other similar choose this or that console based games. I would later get a USB cable that allowed me to download apps and games onto my device. Good times. Decades later and I'm still programming.

I chose .NET and C# for my current startup (2 years in), where I'm the founding lead developer, and I'd still make the same call today. We've been very productive, built a lot of solid features, and two years in we have a stable, high-performing set of apps. We're using ASP.NET Razor Pages, Web API, backend services, and a .NET MAUI Blazor Hybrid app that's published to the app stores - all written in C# using the latest .NET version. Some parts run in Azure, other parts run on high-performance GPU servers on Ubuntu. It has felt like a very practical modern stack, not some legacy-only enterprise choice. Also, hiring has been a non-issue as I had no trouble finding solid C# developers.

Awesome. We've been using Sparrow-0 in our platform since launch, and I'm excited to move to Sparrow-1 over the next few days. Our training and interview pre-screening products rely heavily on Tavus's AI avatars, and this upgrade (based on the video in your blog post) looks like it addresses some real pain points we've run into. Really nice work.

For what it's worth, the startup I currently work for is built entirely in C# and .NET, as was my previous employer. Both startups are based in the Dallas, TX area. Across both companies, applications were hosted on Azure and AWS using a mix of PaaS services and virtual machines running Windows and Linux. We've consistently found this stack to enable strong productivity and high-velocity release cadences.

I've written 3 apps in MAUI with 2 in production.

The 2 apps in production are MAUI Blazor Hybrid apps. The learning curve was very small since I already was familiar with C# and Razor syntax after having built many C# server-side rendered web apps. Development is quite rapid since it's essentially using web technologies. For my use cases, the users do not care about the app looking native as these are B2B apps. Both apps are deployed in the 2 major mobile app stores and 1 app is also deployed in the Microsoft Store.

The 3rd app was an internal-only Android app built with the UI written entirely in C# instead of XAML. I chose to write the UI in C# instead of XAML primarily just for fun and to see how that would work out. I ran into a lot more issues using native UI when it came to me wanting to customize things, such as removing an annoying bottom text underline/border that is added by default to Entry (textbox) controls.

I'll probably never build another native UI app if I have the choice and will stick to MAUI Blazor Hybrid apps because it is so much faster to create the app using web technologies (HTML/CSS).

Great job! And congrats on actually getting something out there. I can really see this being useful for some organizations.

I also envisioned this same type of tool around 10 years ago and it is still on my ever growing list of ideas to implement. I took the idea further to support not only SQL but other languages such as HTML, JavaScript, Python, C#, etc. You could then support returning different types of media based on the URL extension such as .html to return a webpage, .json to return a JSON API, .csv to return a CSV file, etc. As time marched on, many of these same ideas came to fruition in things like AWS Lambda, Jupiter Notebooks, Microsoft Monaco Editor, etc.

I ran an intern program for a year as a software development leader at a previous job. About 25 interns ran through the program that I created. Each intern was paired with a full-time mid to senior-level developer. Out of the interns, only about 5 were decent and required very little handholding while the others required quite a bit of work to answer their questions and keep them on task. Of those 5 decent interns, I only offered full-time employment offers to 2 of them. I believe we paid $16-$20 per hour (the executive team set that rate) and most worked around 10-20 hours per week while a few were able to work upwards of 30-35 hours.

I can only speak for my experience, but I couldn't make it work financially and we would have gotten a lot more done had we used those funds to pay for 1 or more nearshore or offshore resources. It created a fun environment and I think everyone learned and grew from the experience, but from a productivity and financial point of view, I don't think it was worth it.

So based on your experience with Eastern European devs, I think your money would be best spent on hiring 1 or more additional devs from Eastern Europe.

We've been hit by this at work as well. We had to add CAPTCHA and a several other techniques to defend against this.

How it works:

  1.  Attacker leases 1 or more premium rate numbers in an international country.
      - Attacker can lease a premium rate number for as little as $10/month
      - Typically, the attacker gets to keep 70% of the money generated by the premium rate number.

  2. Attacker then finds companies with OTP (One-Time Passcodes) or 2FA (Two-Factor Authentication) endpoints that require no validation and writes a script to automate the webpage or call the API endpoint
     - Attacker will typically obtain a new IP address per API call using a VPN or a rented botnet from the dark web.

  3. If the premium rate number costs 10 cents, then each successful text message they can send to the number generates 7 cents for them.

  4. The attacker then just needs to send 150 SMS to the premium rate number to break-even on their $10 investment, not counting the cost of the VPN or rented botnet.

There is a lot of money to be made here by an attacker unfortunately. :(

Whoa, that's pretty cool and happens to be two of my favorite games :)