eWeekChat

A6: You can't build "real" apps with LC/NC. Some ppl think "real" apps require writing code. It's all about creating value for customer and the business and LC/NC apps can do that.

A6: Myth is: They are only used by low-skilled ‘citizen’ developers or it’s too simple or too messy to achieve value.

In reality, it is possible to create very complex business processes elegantly with no code.

A6: LC/NC apps can't scale. That's not true at all as many of the platforms enable things at broad scale with a broad set of capabilities.

@jonstevenshall So skilled devs are using low code?

(edited)

A6: There are few that are running around. Some are around the quality and maintainability of the generated code compared to human coding. Other myths are around LC/NC breaking cross-team collaboration

A6: LC/NC means no maintenance. That just isn't true as evidenced by some of the conversation that @jonstevenshall started earlier around MS Access.

A6: Low code and no code are in one bucket :) In reality, the user profile, investment, and time to value for the two types of tools are quite different.

A6: Low code/No Code cannot scale and be used in complex apps

@jamesmaguire Yes, skilled devs definitely use LC platforms. Rapid prototyping is an obvious example, but if the platform supports the end goals, LC is a good option for many orgs.

(edited)

@ek121268 there are examples of generated code that works/is maintainable but there are also examples of the opposite. i don't think it's fair to say that's a myth...more that it depends on the particular LC/NC platform/tool. you agree?

@JonStevensHall Good to know. Didn't realize that.

@benlower 100% agree that the tools that are used and their maturity as well as their match to the app that's being developed impacts the overall quality and maintainability

A6: LC requires no ongoing coding or dev ownership. NC requires no set up, integration, or training.

A6: that professional developers dislike low-code platforms. There is a big disconnect between the old generation (let's say the over 45 year olds) and everyone else. https://www.crowdchat.net/s/16a38

source: https://www.digitalocean.com/currents/june-2022

A5: Low code/no code tools provide access to a common layer of supporting services, with AI-driven services being a growth area. #AI services such as clustering & sentiment analysis can be accessed via standard calls to enable higher value innovation by the developer.

AI is so hot right now http://gph.is/1jFLtFf?tc=1 via @giphy

@benlower LOL! And at the same time, true....

A5: I think that the relationship between LC/NC and AI is growing because of expectations. To take the place of a human coder in some cases or expedite his/her work such tool need to be quite intelligent. Tools like https://openai.com/ are an example

A5: I'm not sure that there's a great relationship...yet. There are some examples of ppl using GPT3 on Bubble. But does seem like the LC/NC platforms will add more AI/ML as those solutions become more accessible via well-defined APIs.

@ek121268 OpenAI does a lot of cool stuff.

@JonStevensHall I don't see this

@ek121268 I think #AI capabilities will increasingly differentiate enterprise LC/NC platforms

@benlower I believe that the word relationship is a bit misleading here, and perhaps AI extends LC/NC capabilities not as a dependency but rather as an add-on

(edited)

A5: GitHub Copilot (https://github.com/features/copilot/) comes to mind as an option but is it really "low code" just because it auto programs for you? seems relevant to the LC/NC discussion but might need a new term like "auto code" or something like that.

@LawrenceHecht This is where a lot of investment is happening by platform vendors, particularly in premium enterprise products.

A5: there are a lot of nocode integration platforms that help data engineers with the DataOps workflows. There are many visual "IDE" type tools for aspiring data scientists.

@ek121268 expends or extends? extends makes sense to me if it was a typo. but if not, what do you mean by expends?

@benlower Typo :) Extends

@JonStevensHall a lot of the money is going being wasted. I have a database of vendors in the field.

A5: No direct AI <> LC/NC relationship, but more code faster leads to more access to data lawyers and AI can put this data to good use.

@JonStevensHall I looked at the topic tangentially in this article, which is based on a Kaggle survey: https://thenewstack.io/where-do-data-practitioners...

A5: LC can help share and streamline parts of ML/AI modeling, training, and development across teams.

@LawrenceHecht 5 years ago #AI tended to be a science project, but the current generation of AI libraries and services is industrialised and market-ready. Example: clustering and NLP services detect problem trends in support tickets.

A8: Basics - a lot of NC/LC tools are not teaching the basics of databases. That seems to be one of the hurdles and something that can be fixed.

A8: Low code is not a substitute for an organization's skills, design, or collaboration capabilities. Low code can be a great way to introduce or improve #DevOps practices, removing some of the low-value toil while also encouraging agility, experimentation & teamwork.

A8: Even though it’s LC/NC and value is lowering the bar of technical entry so that biz users can be more directly involved, the fact that these biz users ARE becoming more directly involved will raise their bar of technical knowledge and that's positive investment.

A8: These platforms are being built/updated quite quickly, so make sure to go back to tools that you tested / POCed after every couple of quarters.

A8: did we mention staffing skills, open source and the fact that there are too many platforms? Right now we can't train a workforce on any specific vendor's tools

A8: LC/NC can be competitive advantages if they’re broadly adopted internally with buy-in and used for dev efficiencies and business intelligence.

A8: At the end of the day organizations build products for their clients that must be secure, safe, high performing and high quality as well as accessible. If you properly run a POC for the tools covering all of these requirements you are in a good starting point

@LawrenceHecht there are 100s of platforms, each of them trying to train and certify people.

A8: The need for security is not going away. Democratized development still needs its own set of rules at the company level to ensure compliance.

A7: #AI will be deployed increasingly at the point of development – “co-pilot” assistance will enable more reduction of lower value “toil” & open up new suggestions & insights to enable developers to harness platform capabilities to #innovate & get better results.

@jonstevenshall Sounds like that will certainly come true.

A7: NLP coding would be really cool. “Can we change our getTotal API call to return a number rounded to 2 decimal places?” -> system identifies the API call, finds it in code, finds and returns the transformed value.

A7: I like @jamesmaguire use of co-pilot and I think that such tools will be merged into the tech stack of engineering teams like any other tools. In addition the maturity of the tools will grow more towards autonomous NC generation with greater confidence

@benlower Natural language processing sounds very cool. Code while you're on your Peleton....

haha that would be cool but also terrifying to think we would expect ppl to always be on/productive :)

A7: Demystified, mass adopted, helping close the IT talent gap, considered a basic skill the same way command of MS suite, with also a parallel ecosystem/economy of freelancers that are savvy in many of these tools.

I'm a Zwift person myself... but not when it's 100 degrees farenheit in the UK...

A7: LC platforms and practices will become standard to deal with talent shortage, rate of data growth, and sheer scale of emerging tech, such as AI. NC will be demanded by departments for their teams to capture and act on data.

@JonStevensHall heck, the real point of @OpenAI is to create services developers can use for their apps. For LC/NC, the relevant point is that they create consistent architectures and workflows that can be replicated and audited.