Video 45QmLivYv3k
Key Points
- A McKinsey study reports that developers can finish coding tasks up to twice as fast when using generative AI, especially for repetitive, low‑complexity work.
- Productivity is measured not by lines of code but by delivery-oriented metrics such as DORA (deployment frequency, lead time, MTTR) and project‑management tools like Jira.
- While generative AI markedly boosts team productivity and developer experience, it shows little impact on complex coding tasks, indicating that human developers are still essential.
- AI streamlines workflows through three main capabilities: automating repetitive coding tasks, offering natural‑language interfaces for code generation and debugging, and providing advanced code‑suggestion/autocomplete that helps overcome unfamiliar APIs and “coder’s block.”
- The optimal outcome is a hybrid model where human developers collaborate with generative AI, leveraging these tools to accelerate routine work while retaining human expertise for complex problem‑solving.
Full Transcript
# Video 45QmLivYv3k **Source:** [https://www.youtube.com/watch?v=45QmLivYv3k](https://www.youtube.com/watch?v=45QmLivYv3k) **Duration:** 00:07:47 ## Summary - A McKinsey study reports that developers can finish coding tasks up to twice as fast when using generative AI, especially for repetitive, low‑complexity work. - Productivity is measured not by lines of code but by delivery-oriented metrics such as DORA (deployment frequency, lead time, MTTR) and project‑management tools like Jira. - While generative AI markedly boosts team productivity and developer experience, it shows little impact on complex coding tasks, indicating that human developers are still essential. - AI streamlines workflows through three main capabilities: automating repetitive coding tasks, offering natural‑language interfaces for code generation and debugging, and providing advanced code‑suggestion/autocomplete that helps overcome unfamiliar APIs and “coder’s block.” - The optimal outcome is a hybrid model where human developers collaborate with generative AI, leveraging these tools to accelerate routine work while retaining human expertise for complex problem‑solving. ## Sections - [00:00:00](https://www.youtube.com/watch?v=45QmLivYv3k&t=0s) **AI Doubles Coding Speed, Metrics Explained** - The speaker discusses a McKinsey study claiming generative AI can double developers' coding speed, explores how this might happen, how productivity is measured—including metrics like DORA—and argues AI will boost but not replace human developers. ## Full Transcript
so listen to this a McKenzie study
claims that software developers can
complete coding tasks up to twice as
fast with the help of generative AI
which at least to me brings to mind
three questions and number one
is how how can it do that and we'll get
to that in a moment because I have 10
hows right here secondly I'd really like
to understand how does anybody actually
measure this stuff and anyway how do we
know how productive somebody actually is
and then lastly does this mean that
generative AI could work even faster if
we did away with those human software
developers
entirely well no human developers are
not going anywhere anytime soon humans
we still need you the researchers found
that complex coding tasks were not
significantly affected by the use of
generative AI so worries about AI
replacing developers can be safely laid
to rest for now but generative AI did
drastically speed up team productivity
and improve the developer experience in
a number of lwh hanging fruit use cases
now as for actually measuring developer
productivity so we can tell if
generative AI has helped or hindered
things well you could measure
productivity based on volume how many
lines of code were written but I think
we can all agree that more does doesn't
necessarily mean better I mean if
somebody wants to pay me by the length
of this video I could prattle on for
hours but I don't think any of us would
be feeling particularly productive after
that so fortunately we have metrics
metrics like Dora that stands for devops
research and assessment and that has
metrics like deployment frequency lead
time and meantime to recovery for
evaluating the efficiency of software
delivery and we have project management
tools like jira and those can track
progress manage tasks and facilitate
contribution analysis and what this is
telling us is that productivity Peaks
when we combine human software
developers with generative AI so let's
consider some ways that AI can
streamline development workflows and
number one up here is firstly to
eliminate repetitive task Tas s coding
often involves simple sometimes tedious
tasks and this is where generative AI
tools tend to shine repetitive routine
work like typing out and customizing
standard functions can be expedited
where generative AI simply creates that
code for us and for that to happen we
can use method number two which is the
natural language interface of large
language models so we can ask for the
generation of code snippet FS or debug
code and provide Version Control simply
by asking the AI model nicely in plain
English or whatever our human language
of choice is now number three that is
code suggestion now this goes far beyond
the code assist we've had for years
that's essentially autocomplete but with
code suggestion an AI can help you get
started with an unfamiliar library or an
unfamiliar package or if you're simply
suffering from from coder's block
staring at a blank screen in vs code for
hours well generative AI can get you
going and into the flow of coding sooner
and once you do have some code created
generative AI can then suggest number
four code improvements and it does this
by identifying redundant or inefficient
sections if done right it's like having
a peer programmer sat alongside you
except without the skeptical looks and
number five is code translation now this
is where generative II can translate
code from one language to another
language which could be very useful in
let's say an application modernization
project where we might need to convert
cobal Cod to let's say Java for instance
now I have some more but before we go
into what those are these 6 through 10
let's really consider how generative AI
is able to perform these tasks at all
and it all starts essentially with a
phase called pre trining now pre trining
is what we do with our generative AI
models and we take massive data sets
containing diverse examples of code
written in various programming languages
and we feed it into this pre-training
model the model learns to predict the
the next word or the next token and a
sequence of code based on the context of
the preceding words and this allows the
model to capture the syntax the
semantics and the patterns inherent in
different programming languages now when
it comes to inference time here this is
where we provide a prompt or a query
into the generative AI model and it
processes that input and uses its
learned knowledge to understand the
context and the intent of what we're
asking for now the model considers all
of the relationships between the
different code elements such as the
variables the functions and the control
structures to generate relevant and
syntactically correct code and then
using the Learned patterns and
contextual understanding the generative
AA model can then generate for us an
actual code
snippet that contains the code that we
asked for as the output and this code is
based on the input prompt and follows
the structure and style of the
programming languages which the model
was trained on and this isn't a closed
loop because the model can then adapt to
user feedback helping the model to
refine its understanding and improve
future
outputs okay back to the way that
generative AI can help with coding now
where did we get to yeah number six so
number six that is code testing now it
can create test cases by analyzing code
and generating test inputs and then even
run those inputs and evaluate the
resulting outputs in a similar vein AI
can number seven perform bug detection
in identifying and even automatic fixing
bugs AI can also help create
personalized Dev environments that adapt
to individual developer preferences and
coding Styles and in a personal favorite
AI can be used to generate
documentation so I don't have to by
summarizing code functionalities
explaining algorithms and providing
context look well generative AI encoding
is a powerful tool it is not a
substitute for the creativity problem
solving and domain expertise of human
developers we should think of it as an
augmentation tool something that assists
developers in coding tasks providing
suggestions and potentially speeding up
certain aspects of the development
process oh and no no no I haven't
forgotten haven't forgotten number
10 because that one's for you how has
generative AI improved your coding
experience is there something I haven't
mentioned here let me know in the
comments
if you like this video and want to see
more like it please like And
subscribe if you have any questions or
want to share your thoughts about this
topic please leave a comment below to
learn more please reach out to your IBM
sales team or IBM business partner