Designing The Future Of Software Engineering with Amjad Masad from Replit

[@ARKInvest2015]

ARK is aware of hacked third-party KZbin channels fraudulently posing as ARK &/or Cathie Wood. These accounts are impersonators & not affiliated with ARK Invest or Cathie Wood in any way. ARK Invest will never use KZbin or other social media to solicit money, including cryptocurrency. Please report any account claiming otherwise! This channel is the only ARK Invest KZbin channel. For more information, please view our "What Types of Scams Impersonate ARK Invest and Cathie Wood?" section on our FAQ page here: ark-invest.com/faq/

[@stokes7718]

Thanks for this. I don't see this is as transformational. I can easily believe that this will accelerate the creation of prototype systems. I can also see that the way replit solves programming environment challenges is a big improvement. But, most of the cost of software development is in design, testing, maintenance and support, Writing code is about 20% of the cost. Even if the generated code is easy to maintain I don't see any way to get more than a 25% improvement the full lifecycle costs from technology like this. Having said that a 25% cost reduction is well worth having.

[@crossfitglenburnie2221]

AI assisted code development may result in fewer errors delivered to subsequent phases. Cycle time may improve as the coding time reduces. And, who knows, testing-as-data may improve development, if it too becomes part of the AI instrumentation process.

[@stokes7718]

@@crossfitglenburnie2221 Thanks for taking the time to reply. You are probably right about there being fewer errors and faster cycle times. That's why I think the overall improvement is likely to be greater than that you would get by reducing the coding cost to zero. I also agree that testing-as-data would be a real game changer if it could be done.

[@crossfitglenburnie2221]

@Justin Rand the decry of software commoditization has been heard for decades. Ed Youdon published a book back in 1993 titled "Decline & Fall of the American Programmer". Yourdan's argument - in part - was that the willingness of Indian programmers to use case tools would enable them to produce good, cheap software at a fraction of the cost of that generated by American 'cowboys'. What we saw over the decades was actually a move toward more efficient and agile development. These efficiencies - gained through agile processes, CASE tool advancement, low-code development, and innovation in general - would yeild more and more complex software systems. Parkinson's Law may apply - where work expands so as to fill the time available for its completion - first observed as the growth of bureaucracies over time. It has since been applied to software complexity, and there are corollaries which suggest that increasing software complexity in fact enables bureaucratic growth. So, I would posit that complexity is the problem, and the value of AI may be best rooted in its affinity to see more clearly (than humans) in that complex environment.

[@johnheld8770]

I think the approach to coding is going to change. A lot of code will be generated from speech that is closer to natural language. Things like RPA experience even more rapid growth. The value of testing skills will increase as LLM generated code will sometimes exhibit unanticipated behavior.

[@VideoOverdrive]

Is Ark invested in Replit?