These aren’t catastrophic numbers. In fact, for many tasks they’re impressive. But they do clarify something important about where human judgment fits in the new workflow.

In a post we published earlier this month, we talked about the shift from vibe coding to agentic engineering, a term Andrej Karpathy, who coined “vibe coding” in the first place, now prefers.

The distinction matters.

Vibe coding was prompt-and-hope. Agentic engineering means orchestrating AI systems with real oversight. You define the architecture, set the quality gates, and own what ships.

The Waterloo research puts a number on why that oversight layer isn’t optional. If AI tools are wrong 25% of the time on simpler tasks—and more often on complex ones like image, video, and website generation—then the engineer’s job isn’t disappearing. It’s changing.

Their job is increasingly to catch the 25%.

That requires seniority. It requires knowing what good architecture looks like before the agent starts writing. And it’s why we keep coming back to the same principle—AI augments, it doesn’t replace. Every line that ships is still a human decision.

This is especially relevant if you’re evaluating an outsourced development partner right now. The question isn’t whether they use AI tools. Anyone worth working with does. The question is what their human layer looks like around those tools.

If you want to talk about keeping humans in the loop as you adopt AI, I’m always available for a call.

— Wes

From our blog and around Waterloo Region

• Where to cowork in Waterloo Region

• Give to Gain: Inside Communitech’s International Women’s Day | Tech x Women Event

• AI bridging the gap between health-care visits

In that moment, I realized something that no spreadsheet or pitch deck could ever teach me: Trust is the most fragile currency in business.

In the tech world, it’s incredibly easy to get distracted. We talk about features, obsess over tech stacks, and chase the latest AI “bullseye.” But at the end of the day, business doesn’t happen between machines. It happens when two people look at each other and decide they can rely on one another.

When you bring on a software development partner like BitBakery, you aren’t just buying lines of code. You are handing over the keys to your product, your customer experience, and your reputation.

That is a massive weight. And frankly, the traditional “vendor” model isn’t built to carry it.

A vendor follows a spec sheet. A partner owns the outcome.

In last month’s newsletter, we talked about why it’s time to stop “adopting” AI and start adapting to it. The same logic applies to how we build software. Adaptation requires a fundamental shift in how we work together.

At BitBakery, we’ve built our culture around a few simple, non-negotiable pillars:

• We do what we say we’ll do. Reliable delivery is the baseline, not the goal.

• We communicate clearly. No jargon, no “black box” development, just honest updates.

• We own outcomes, not excuses. If there’s a hurdle, we solve it together.

Trust isn’t something you can write into a tagline or buy with a marketing budget. It’s earned—one delivery, one deadline, and one honest conversation at a time.

If you’re tired of the vendor runaround and looking for a software partner you can actually rely on, we’re here.

Let’s build something real — together.

— Wes

From our blog and around Waterloo Region

• $7.5 million seed funding helps Upside Robotics plant the seeds for success (Velocity)

• Apple strikes deal to acquire Canadian database software startup Kuzu

• Waterloo Region’s Moonshot: Where Founders & Physicians Redefine Medicine (Communitech)

So what did I think about? The same thing you’re probably thinking about now after scrolling through LinkedIn or reading the news. That’s right. AI. We might have taken a break, but the pundits and promoters of AI did not. We’re only a few days into the new year, and already we’re seeing more predictions around AI adoption, AI bubbles, and what to do about AI slop.

It can be easy to think of AI as a recent paradigm shift. OpenAI released ChatGPT in November 2023, and it feels like it’s the only thing we’ve talked about for the last two years. In that short time, we’ve seen a rush to adopt AI—sometimes without taking the time to truly understand the problem we’re trying to solve and what the impact AI will have on our businesses and teams.

It’s the return of the “move fast and break things” mentality on an exponentially larger scale than any of us could have imagined.

But AI isn’t new. Nor our efforts to adopt it into how we work.

AI pioneer Jeffery Hinton began his research at the University of Edinburgh in the 1970s. Advancements in machine learning and computer vision over the last four decades have given us everyday technologies, from face and fingerprint scanners to recommendations on Netflix and Spotify.

A locally-connected story over the holiday break reminded me of how long AI has been around. AI chip developer Groq (not to be confused with Elon Musk’s Grok) announced a licensing deal with Nvidia to the tune of $20 billion. One of the early investors in Groq was Garage Capital, founded by local tech leaders Michael Litt, Mike McCauley, and Devon Galloway.

It was an investment made ten years ago that has paid off handsomely for Garage Capital’s partners. It’s also a reminder that adopting AI isn’t meant to happen overnight. Leaders who understand that successfully planning AI adoption is a long game are the ones who will reap the rewards.

If you want to talk about AI adoption and the challenges your business is facing, I’m always available for a call.

- Wes

From our blog and around Waterloo Region

• Groq’s staggering Nvidia deal has strong Canadian connections

• Two Canadian hospitals giving startup’s imaging tech a chance

• Lessons from Trusscore’s CEO on why exits aren’t endpoints

We’ve learned a lot about the capabilities (and limits) of artificial intelligence over the last twelve months. Our team has experimented with almost every tool, from Claude to Gemini to Lovable. We’ve shared how we use AI for pair programming, how AI is transforming search, and how AI is changing the software development life cycle.

With all the advancements in generative AI and agentic AI, we can all agree on one thing:

AI is powerful, but without a human at the helm, a project can quickly fall off the tracks.

That’s why our New Year’s resolution is to continue pairing human expertise with AI experimentation and learning to deliver best-in-class solutions for our clients.

But before we log off for a much-deserved break, I wanted to share some thoughts on what we’ve accomplished in 2025 and what we’re excited about in the year to come.

This time of year offers a moment to appreciate the people who have trusted us with their ideas, the partners who have challenged us to think differently, and the teams who continue to show up with curiosity and generosity.

BitBakery would not be BitBakery without that mix. Here’s what we’re celebrating this season:

• A year of meaningful builds. From agritourism to insurance to digital therapy tools, our team delivered work that truly matters. That has always been our measure of success.

• The rise of thoughtful AI. 2025 was the year AI shifted from “what if” to “how do we do this responsibly.” We are proud to have helped clients navigate that change with clarity instead of hype.

• Our team keeps levelling up. The talent inside BitBakery continues to impress, and the kindness matches the capability. Not every company gets both.

No one can predict the future, but here are three things I’m excited about exploring in 2026:

• Smarter products, fewer silos. Next year, our work will focus on integrating content, data, systems, and teams so the products we build feel seamless for every user.

• AI with a purpose. Not AI for attention. Not AI to check a box. AI that actually solves problems, removes friction, and adds real value. You can expect more experimentation from us in this space, particularly in healthcare and education.

• Doubling down on experience. The bar for strong UX keeps rising, and we welcome it. In 2026, there will be a deeper focus on accessibility, mobile-first design, and inclusive digital experiences.

Whether you partnered with us, followed along, or simply read these newsletters each month, thank you.

Building things is what we do. Building them with people like you is what makes it meaningful.

Wishing you a restful and joyful holiday season,

- Wes and the BitBakery team

From our blog and around Waterloo Region

• BitBakery: What you need to know about coding with an AI assistant

• BetaKit: Assetflo closes $2 million USD to make clients’ every physical asset “searchable”

• BetaKit: Nfinite nabs $4.6 million to prove it can make paper that performs like plastic

• AC:Incubate Spotlight: Qasim Parekh, co-founder of CellaraRx

Creation by mere thought.

It’s a science fiction concept that suddenly feels real thanks to generative AI and the rise of vibe coding. Tools like Lovable, Base44, Cursor, and Anthropic’s Claude Code can turn prompts into web and mobile apps in minutes. Our team has been experimenting with these tools to see how they can improve our processes and deliver the same quality that our clients expect.

But it’s not just developers who are vibe coding new projects. Product managers, designers, and almost anyone can use one of these tools to build an app.

Some have called it a democratization of software development. Anyone can code. But that doesn’t mean the code is production-ready.

Software development is more than just functions, arrays, and APIs. We’ve seen a number of vibe-coded projects that look great from afar, but once we dig into them, we see the vibes are off.

One of the BitBakers referenced a movie I haven’t seen, but have added to my list. Riffing on Cher, the main character from Clueless, they described most of these vibe-code apps as “being full-on monets”.

They’re great from far away, but once you dig into them, there is still much work to be done.

So what’s missing? Vibe code projects can be built quickly, but they often overlook long-term maintainability, scalability, and comprehensive testing.

When our team uses generative AI for development, we follow three core principles:

1. AI augments, not replaces. We believe AI should empower people — not erase them.

2. Humans own every line of shipped code. AI can assist, but accountability always stays human.

3. Every use of AI must be driven by clarity, curiosity, and critical thinking. We don’t chase hype. We build with intent.

In the rush to define what’s possible, we think it’s just as important to define what’s right.

- Wes

From our blog and around Waterloo Region

• BitBakery: From Ideas to Impact: Innovation at KidsAbility

• BitBakery: What OpenAI’s 2025 DevDay Means for Developers

• Medical Innovation Xchange: MIX Profile: Rachel Bartholomew, founder and CEO of Hyivy

• Communitech: Community and tech take centre stage at Techtoberfest 2025

• Velocity: Forechecking hard against undetected concussions

• Accelerator Centre: AC:Incubate Spotlight: Salar Enjedani, Founder and CEO of Innorobix Automation

AI is transforming the way we work in so many ways, from searching and researching to coding and creating. But there’s one concept that AI is changing that I think doesn’t receive the attention it deserves.

Writing a prompt to generate an output was a novel concept when OpenAI released the first public version of ChatGPT in 2023. Today, almost everyone is comfortable crafting a prompt as if it were second nature. But when I speak with people, the conversation is on the output, not the input.

Here’s my thought. The prompt is the new user interface. While Apple and Google work on updated design systems for their aging operating systems, OpenAI (and the majority of other LLM makers) have realized that the prompt is how we interact with their systems.

This idea was reinforced for me yesterday during OpenAI’s DevDay 2025 keynote. OpenAI CEO Sam Altman introduced its new Apps SDK that lets developers build full-fledged, interactive web apps that run inside ChatGPT conversations. Not only is the prompt the interface, but now ChatGPT is the operating system.

It reminded me of one of my favourite scenes from “Seinfeld”.

Kramer gets a new phone number, which is one digit off from the Moviefone number. He’s got time, so he starts answering the calls and fielding inquiries from his fellow New Yorkers. But it was the call with George that made me think of LLMs and chatbots. Unable to decipher what button George is pressing, Kramer is exasperated and says, “Why don’t you just tell me the name of the movie you’ve selected?”

Looking back at that scene today, you can see the seeds of how we interact with LLMs today. Or as Jerry put it, “Oh I gotta get on that internet, I’m late on everything!”

Prompts are becoming the interface paradigm that is changing how we design, build, and collaborate. If you’ve tried (or want to try) a ‘prompt-as-UI’ workflow, reply with what you’re experimenting with. I’ll send back suggestions and share a few examples next issue.

- Wes

From our blog and around Waterloo Region

• BitBakery: Can AI replace pair programming?

• BitBakery: What OpenAI’s 2025 DevDay Means for Developers

• Medical Innovation Xchange: MIX Profile: Michael Phillips, Co-founder and CEO of Vena Medical

• The Record: Waterloo councillors invest in med-tech sector

• Communitech: ChatGPT joins the conversation as Waterloo Region talks AI and quantum

• Communitech: How MissionSync helps teams close the gap between plans and results

• Communitech: Alchemy secures $6M to scale global defence & automotive innovations

• Velocity: Waterloo founders raise more than $14 billion USD in 2024

• BetaKit: FluidAI lands FDA clearance to sell post-surgery monitor in the US

• Accelerator Centre: Founder Spotlight: Majid Mirza, Founder and CEO of ESGTree

Thanks for reading BitBakery's Baker’s Dozen! Subscribe for free to receive new posts and support my work.

There aren’t many experiences quite like launching a new software solution. But when the confetti has been swept up, what happens after launch day?

Software issues don’t always show up on launch day. Sometimes, they can take weeks or months to appear. Most teams plan the build; fewer plan the care. If DevOps sounds like “future us will deal with it,” this one’s for you. We've put together a quick post on how we approach DevOps, highlighting six lessons that help keep releases calm and predictable.

Our model wraps support and maintenance into the software development lifecycle. That means clear SLAs, real incident playbooks, and post-mortems that actually drive changes in the next sprint. You get steady velocity without burning your people or your budget.

If you’re planning Q4 features or hardening for 2026, we can help you stand up CI/CD, observability, and incident workflows that stick. Hit reply and tell us what “success” looks like over the next 6–12 months, and we’ll map the path.

- Wes

AI corner: GPT-5 and open weights (what devs should know)

OpenAI shipped GPT-5 and new open-weight models (gpt-oss-120b/20b) last month. They’re both useful for teams that need control over infrastructure, data, and cost. We broke down where GPT-5 helps (reasoning, coding, agentic tasks) and when open weights make sense.

Waterloo Region tech headlines

• Astrus announces $11 million CAD to automate semiconductor design with AI (BetaKit)

• Bankruptcy docs claim faltering sales pipeline sunk Uvaro and Lighthouse Labs post acquisition (BetaKit)

• BinSentry raises $68.8-million CAD Series C as it brings remote animal feed monitoring platform to Brazil (BetaKit)

• Waterloo’s Plum earns top HR tech award (Communitech)

• Friendlier’s total funding tops $10M with new $4.5M raise (Communitech)

• Reimagining entrepreneurship through equity and inclusion (Accelerator Centre)

• MIX Profile: Sydney Robinson, Co-founder and CEO of Vessl Prosthetics (Medical Innovation Xchange)

Coworking hubs have become an integral part of our innovation ecosystem, offering more than desks and Wi-Fi. They’re places that foster community, connection, and those serendipitous moments that spark the next great idea.

(P.S. We recently rounded up our favourite local coworking spots on the blog, if you’d like a deeper dive.)

BitBakery’s workspace journey

Our workspace story has evolved from a heads-down private HQ that offered focus but little outside interaction, to co-locating with partners, before finally landing at Communitech, a purpose-built community of founders, mentors and industry experts in the heart of Downtown Kitchener.

Each stage taught us something new, but coworking has proven the best match for our hybrid model and growth ambitions.

Why Coworking Works for Us

• Flexibility – We can scale desks up or down as the team shifts between home and office.

• Amenities – Call booths, meeting rooms, a gym, even an indoor basketball court, turn a workday into an experience.

• Community – Daily proximity to entrepreneurs, investors, and innovators leads to ideas and partnerships we couldn’t manufacture alone.

• On-site Support – Communitech programs and advisors sit just steps away, shortening the distance between a question and an answer.

Looking Ahead

Hybrid work isn’t a fad, it’s the new normal. Coworking hits the sweet spot between a professional environment and the freedom to choose where (and how) we work. For BitBakery, the space itself is now an accelerator of connection and creativity.

Closing Thought

Whether you’re a freelancer looking for camaraderie, a startup hunting for its first office, or an established team craving fresh perspective, coworking could be your next smart move.

Happy International Coworking Day from the BitBakery crew. See you around the hot desks!

- Wes

Waterloo Region tech headlines

• What developers need to know about GPT-5 and GPT-OSS

• What businesses need to know about AI search

• Cybersecurity startup Cavelo secures $5-million seed extension round led by Inovia Capital

• eleven-x celebrates 11 years in business and lands its biggest deals yet

• GeoMate partners with City of Kitchener

Video games have hundreds of blog posts with tips, tricks, and advice. The same challenges exist for software developers, but over the last few months, developers have been turning to generative AI tools to help overcome seemingly impossible bugs.

A developer on our team recently described using an AI tool for help with a particularly ugly bug. In the time it took them to sip their coffee, the model spit back a multi-step patch, unit tests, and a commit message.

“My eyes hadn’t even scanned the stack-trace* yet, and the AI had already resolved it.”

While our developer was still considering where to start, the AI had already figured it out and provided a working solution.

The mismatch between human processing speed and AI output left him feeling oddly out of sync, both impressed and mentally overloaded. Industry reporters are starting to label this phenomenon “AI burnout,” noting that the triple load of context switching, hyper-speed iteration, and constant vigilance is uniquely taxing for developers.

Remember the early-pandemic Zoom marathons?

Think back to spring 2020, when our calendars turned into a series of back-to-back video calls and we couldn’t wait to slam our laptop lids shut.

Stanford researchers later unpacked why those days left us exhausted. Enlarged faces triggered fight-or-flight responses—we were looking at ourselves nonstop, stuck in our seats and performing on camera.

AI tools trigger a similar overload, just in reverse. Instead of too many of us on screen, we’re staring at a stream of machine-generated answers that arrive faster than we can evaluate them.

The result is the same: mental fatigue, decision paralysis, and a creeping sense that we’re the bottleneck in our own workflow.

Why does this cognitive dissonance happen?

When a tool’s velocity eclipses our cognitive pace, every interaction feels like sprinting to keep up. Over time, that tension accumulates as stress, second-guessing, or the urge to double-check everything.

3 ways to demystify the machine

1. Walk with a guide. Partner with a team that’s been down the road (that’s us). We help teams design workflows where humans remain in charge—from prompt engineering playbooks to human-in-the-loop review cycles—so you gain speed without surrendering comprehension.

2. Sprint before you marathon. Scope a two-week proof-of-concept that answers one business question. Small bets surface integration snags, ownership gaps, and governance issues early, saving you from enterprise-wide refactors later.

3. Build “mental margins” into your AI stack. Adopt meeting-hygiene habits for tooling:

   • Time-box prompting sessions the way you schedule meetings

   • Insert reflection checkpoints. After every major AI suggestion, require one human validation pass before merging or publishing

   • Rotate “driver” roles so no teammate is on continuous prompt duty. These micro-buffers slow the process just enough for your cognition to catch up, preventing the lag that fuels fatigue

If AI velocity has you feeling like you’ve skipped a beat, let’s talk. BitBakery’s AI Enablement Team turns “whoa, that was fast” into “wow, that was manageable.”

Hit 'Reply' or book a 30-minute discovery call. Your brain will thank you.

Until next time, keep your prompts sharp and your coffee warm.

* A stack-trace is like the program’s breadcrumb trail: when something goes wrong, it lists each function or method the code stepped through until the error occurred. By showing the exact line numbers and call order, it helps developers pinpoint where the bug surfaced and what led up to it, much like retracing your route on a map to find where you took a wrong turn.

- Wes

Waterloo Region tech headlines

• ENVGO raises $2.7-million seed round to bring its electric “flying” boat to market

• Bridgit’s Mallorie Brodie explains why Canadian startups need to think globally from the start

• Nicoya acquires UK’s Applied Photophysics to expand its drug development capabilities and global reach

• Accelerator Centre Client Spotlight: Stefanie Bruinsma, Founder of AutoCate

One thing became clear. AI might be everywhere, but genuine understanding of its practical use (and its limitations) is still evolving. Many pitches claim revolutionary AI solutions that are just improved workflows.

Beyond the hype, three critical conversations stood out at WebSummit. Call them the three S’s—security, scalability, and semantic understanding.

The first two are familiar challenges, but semantics is the one that’s quietly becoming essential. AI tools today still struggle to grasp the real-world context and the nuanced meaning that humans intuitively understand. That context gap means AI isn't yet ready to handle complex projects independently.

This gap echoed clearly during our recent BitBakery AI hackathon. We experimented with a variety of new tools that helped us quickly build usable prototypes.

We also discovered firsthand that AI requires skilled human developers and designers who have a deep understanding of users and their context. AI might generate great ideas and initial frameworks, but it takes human insight to turn them into genuinely impactful solutions.

AI is rapidly advancing, but real breakthroughs will depend on how well we bridge the semantic gap. I'm excited to keep exploring this balance at BitBakery and to share what we learn with you.

What gaps or opportunities are you overlooking in your AI journey? Let's talk.

- Wes

Waterloo Region tech headlines

• Why Faire’s co-founder wants to bring Silicon Valley’s DNA to Canada

• University of Waterloo-backed Velocity Fund II holds $14-million initial close

• Serving up a taste of robot food delivery

But last week, our team tried something new. We went on a team paddle with Canoeing the Grand. And to our surprise, none of us had ever done it before. It was one of those quiet, perspective-shifting moments—how many experiences are hiding in plain sight, just waiting for us to notice?

That got us wondering about what else we are overlooking.

Not just around town, but at work. What habits have we settled into that keep us from discovering something better? A different way to reflect on a project. A fresh approach to team-building. A tool we’ve scrolled past a dozen times but never actually tried.

This week, we explored one of those tools on our blog. GitHub’s Copilot Agent Mode. You can read what we learned about the benefits and risks of AI agents on our blog, along with another new post from BitBaker Annaliese Schropp on how she uses ChatGPT with Visual Studio Code.

Speaking of trying new things, I’ll attend two new conferences this May. If you’re interested in meeting up at Consensus 2025 (May 14–16) or Web Summit Vancouver (May 27–30), please send me an email or respond to this newsletter.

As always, I'm always happy to have a call if you have questions about AI or other development and design questions.

- Wes

Waterloo Region tech headlines

• Communitech appoints Sheldon McCormick as new CEO following layoffs (BetaKit)

• Dozr founding CEO Kevin Forestell moves to chair while Dave Frazier joins as president (BetaKit)

• ConeLabs turned a pitch win in Waterloo Region into real-world government work (Communitech)

• AC:Studio Innovation Showcase Recap (Accelerator Centre)

• GRT presents its 10-year plan (The Community Edition)

Personally, I have always found that making time for reflection helps me stay focused on the challenges and opportunities around me. I set aside time each day to journal and set goals for the days and weeks ahead. Journaling and self-reflection were also two of the lessons we learned from resilience training sessions with Serena Jain.

The team does this as a group in our Knowledge Nibbles lunch and learn sessions. Last month, we got together to share how we use artificial intelligence (AI) in our work and personal lives. It was interesting to hear each BitBaker approaches AI—from using it as a sounding board for development questions to creating closed captions for videos. You can read all the ways we use AI on the blog in the post “How AI can help boost your productivity”.

The new year is also a time for starting new projects. If one of those new projects is bringing a business idea to life, then Waterloo Region has the ingredients you need to make it happen. We’ve put together a guide to all of the local resources available to help founders take their vision from idea to reality in the post “Where to get help for your startup in Waterloo Region”.

As always, if you have questions about AI or other development and design questions, I’m always happy to have a call.

- Wes

Waterloo Region tech headlines

• Blockchain entrepreneur Sergey Gorbunov donated $1 million to the University of Waterloo to build a new lab. The funding will also support fellowships for 12 PhD students and six undergrads.

• Communitech celebrated the first group of graduates from its Startup Bootcamp program, supported by the City of Kitchener.

• VueReal secured a $58.4-million CAD Series C to boost its semiconductor production.

• The Federal government announced a $9 million investment in the Accelerator Centre’s AC:Studio program.

We did just that last week at the newly renovated and reopened Boathouse in Downtown Kitchener’s Victoria Park. The venue has a storied history as one of the city’s best live music venues. The group behind the new Boathouse has built on that history with an eye on the future of arts and culture here in Kitchener.

The murals in the venue showcase that history well, and it made me think of something that pops up in conversation around the office. Whether it’s a live music venue or an insurance company, having access to your historical information is crucial to delivering on your future plans. Unfortunately, that institutional knowledge can be lost when people leave for other opportunities. You may have terabytes of notes, presentations, and spreadsheets—but can you find what you need when you need it?

Some people see artificial intelligence as a tool to generate more content, but what if we looked at AI as a way to access the institutional knowledge hidden in our Google Drives or Sharepoints?

Reflecting on the importance of preserving knowledge and building a strong foundation for the future reminds me of the value of connection. The holidays are a time to celebrate those connections and express gratitude for the people who make everything possible.

From all of us at BitBakery, we wish you a joyful holiday season filled with warmth, reflection, and inspiration for the year ahead.

- Wes

Our top five posts from 2024…

• Where to cowork in Waterloo Region

• Creating icon libraries with ChatGPT

• The BitBakery guide to baked goods in Waterloo Region

• How to Use Apple HealthKit Data in an App

• Is the Arc browser the ultimate tool for developers?

AI hype presents many possibilities, but I’ve found product leaders and business owners are struggling to make generative AI work for their needs.

If you scroll LinkedIn or Threads, you’ll inevitably see posts with “experts” sharing five prompt writing tips you need to know today. While prompt writing is an important emerging skill, it only touches the surface of what these AI platforms can do.

This inspired us to write a blog post about GPTs, a tool in most generative AI platforms that lets you train specific skills and knowledge into a reusable AI app.

If you’re experiencing challenges with prompt writing or creating GPTs, I’d like to see if we can help.

- Wes

From the BitBakery Blog…

• We recently moved our website from Statamic to Webflow. Attila Schmidt shares what we learned and what you need to know to choose the right web platform for your business.

• We're exploring AI opportunities for our customers and finding ways to use AI in our internal workflows, too. Here’s a video I recorded on how we’re doing it.

• Are you looking for a way to simplify cloud infrastructure deployments? Annaliese Schropp shares her experience taking over a project that uses Terraform to manage infrastructure as code (IaC) and bring automation and repeatability to setting up cloud environments.

The hype surrounding generative AI tools like ChatGPT, Midjourney, and Microsoft Copilot is something to behold (see the chart above). But that hype doesn’t always translate to success for companies adding AI features to their applications.

Mike Kirkup has some thoughts on this. Mike is the CTO of EnPowered and Arlo and a former colleague from UW Velocity. He was on a recent episode of Communitech’s Tech About Town podcast discussing fundraising with MaRS Discovery District’s Vartika Manasvi when the topic of startups adding AI features came up.

“I think the part that we're getting frustrated with is that people aren't thinking about what to use AI for. They're like, ‘If I could just slap AI onto this, then it's better.’ And the market is correctly saying, ‘I'm not going to pay for that.’ Slack spent a ton of money adding AI functionality. Most people aren't paying for it. Salesforce added a ton of sale AI functionality. Most people aren't paying for it. Notion they wanted to charge extra for AI now they've just rolled it into the main product.”

I’ve known Mike for a long time, and he understands what markets are genuinely looking for. He says that the real value of AI is not in additional features but in native AI applications and services. He gave an example that illustrates this point well.

“A really great analog is GPS. There were very clear examples of apps that couldn't exist on mobile before GPS. Turn by turn navigation was not possible. You could have a map. It could tell you, I want to go from A to B, but where are you? What's missing was enough that you would carry around an actual portable GPS puck that would tell you where you are. Then we put it in the phone and Uber comes along and completely redefines how you get around in a car, because they know where you are and where you want to go, where the driver is, and this whole experience was amazing. No one ever thought of it like that. That still has to happen with AI, and I think that's the frustration by consumers right now is that AI is only making an experience 10% better. We just have to be patient.”

Have you had a native AI experience that has wowed you? I’d like to hear more about it. 

- Wes

From the BitBakery Blog…

• Notifications can make or break your app experience. In her latest blog post, BitBaker Annaliese Schropp explores building better notifications that engage users without overwhelming them.
  

• Android 15 and iOS 18 are out in the wild. Attila Schmidt unpacks what these updates mean for your business and how what opportunities for new features these releases offer.

It reminds me of one of my favourite movie quotes:

“Life moves pretty fast. If you don’t stop and look around once in a while, you could miss it.” – Ferris Bueller

Like any technology, it’s critical to understand the benefits and risks of implementing generative AI or other AI technologies into a solution. We put together a blog post on one aspect of this—AI guardrails. You can find the link below. We also asked the team what they think people are getting wrong about AI—here’s what they said:

I’d love to hear what you think.

- Wes

From the BitBakery Blog…

• The upcoming releases of Google’s Android 15 and Apple’s iOS 18 will bring smartphone users several new features, tools, and AI-powered enhancements. Here’s what to expect in Android 15 and iOS 18.
  

• If your business is using or planning to implement generative AI in your products or workflows, ensuring that the answers or data being generated are as accurate as possible is critical to avoid reputational and legal issues. Our team put together a blog post on what you need to know about AI guardrails.

On May 13, OpenAI released Chat GPT 4o, a major update to its generative AI tool. That wasn’t all. They also made it available (with some usage restrictions) to all users.

Then, on June 10, Apple finally announced its AI plans with the beautifully named “Apple Intelligence” at the Apple World Wide Developer Conference.

Coming in future iOS, iPadOS, and macOS updates, Apple claims its AI tools will deliver a new level of functionality for the devices in our pockets while still protecting user privacy. Part of this update includes a deep integration with OpenAI.

One thing that really stood out to me during the Apple presentation was the company’s focus on context as being critical to making generative AI work well. When you send the right context with a prompt, you’ll get better results. 

We’ve put this into action with a no-prompt engineering approach in our GhostWriter AI Pro application. Instead of getting lost writing the perfect prompt, no-prompt lets you set your tone and output length. 

We put together a quick demonstration of this using the latest versions of Microsoft Copilot AI, OpenAI ChatGPT, Google Gemini, and GhostWriter AI Pro.

We generated an email using the same prompt in each tool. 

The prompt: “Can you write an email asking a client for the information needed to complete the website update project? We're looking for the desired deployment date and if they have a window where we can ensure users aren't adding in new data. We can't proceed without this information. Need to be friendly, but assertive.”

Here are the results:

Gemini

Subject: Quick Update Needed to Finalize Your Website Refresh!

Hi [Client Name],

I'm reaching out to finalize a few details to ensure a smooth launch for your website refresh!

To wrap things up efficiently, we'd love to know your ideal deployment date. This helps us plan our development timeline and keep you informed on progress.

Additionally, for a seamless transition, it would be helpful to know if there's a specific window when minimal user activity is expected. This allows us to schedule the update during a downtime period to minimize disruption to users adding new data.

We understand you might be busy, but having this information is crucial to move forward with the project. Please let us know your preferred deployment date and any downtime considerations at your earliest convenience.

In the meantime, feel free to reach out if you have any questions or need further clarification!

Thanks,

[Your Name]

P.S. We're excited to bring your website vision to life!

ChatGPT

Subject: Important Information Needed to Complete Your Website Update

Hi [Client's Name],

I hope this message finds you well!

We're excited to move forward with your website update project, but we need a bit more information from you to ensure everything goes smoothly. Specifically, we're looking for:

1. Desired Deployment Date: When would you like us to deploy the updated website? Having this date will help us align our schedules and ensure we meet your expectations.

2. Data Entry Window: Is there a specific window of time when we can perform the deployment to ensure users aren't adding new data? This will help us avoid any potential data inconsistencies or loss during the update process.

Your prompt response to these questions is crucial for us to proceed without delays. We want to ensure the update is seamless and meets all your requirements.

Thank you for your cooperation! Looking forward to your response.

Best regards,

[Your Name]
[Your Position]
[Your Contact Information]
[Your Company]

Copilot AI

Subject: Request for Deployment Details for Website Update Project

Dear [Client’s Name],

I hope this email finds you well. As we move forward with the website update project, we require some essential information to ensure a smooth deployment process. Your prompt response will greatly assist us in meeting our project deadlines.

Desired Deployment Date: Could you please provide us with the preferred date for deploying the updated website? This will allow us to schedule our resources accordingly and plan for a successful launch.

Data-Free Window: Additionally, we’d like to know if there is a specific window during which users won’t be actively adding new data to the system. Having this information will help us minimize any potential disruptions during the update.

We understand the importance of timely communication and appreciate your cooperation. If you have any questions or concerns, feel free to reach out to us.

Thank you for your attention to this matter, and we look forward to receiving the requested details.

Best regards,

[Your Name] [Your Position] [Your Contact Information]

Feel free to customize this email to fit your specific context and client relationship. Good luck with your project! 😊

GhostWriter AI Pro

Subject: Request for Project Details Needed for Website Update

Hi [Client’s Name],

Hope you’re well! To keep the train on the tracks for your website update, we need two key pieces of info:

1. Desired deployment date – when should the magic happen?

2. A time window to ensure users don’t add new data during updates – we need absolute focus here!

Without these details, our hands are tied, and we can't move forward. Could you please provide this information at your earliest convenience? We’re eager to get rolling!

Best,

[Your Name]

The results are interesting, and I’d like to know which result you think came out the best.

- Wes

From the BitBakery Blog…

• You’re reading a newsletter, which means you might already know how powerful they can be for building a community. In this post, we dig into what you need to know to create a newsletter that helps you connect with your users and grow your audience.

• We’ve got a new case study on how we used an embedded team approach to help a client deliver a new web application. In this example, we provided a front-end design and development team that allowed the client’s team to focus on its backend work.

Every episode provides part of an answer to an overarching question and then adds ten new mysteries before the credits begin to roll. Generative AI tools like ChatGTP and Co-Pilot remind me of a season four episode of ABC’s LOST. It’s been a while, but in the episode, the show’s villain (or anti-hero, depending on your viewpoint) describes a box that can bring anything you think about to the mysterious island.

Like most of the series, there’s no explanation for how this works—which is very similar to generative AI. We write prompts and it gives us what we asked for, but we have very little insight into how that “magic” happens.

Last summer, we held an internal ChatGPT hackathon to dedicate time to seeing behind the generative AI curtain. Our team learned a great deal about prompts, how they’re structured, and how they influence a generative AI’s output.

We used those learnings to turn one of the hackathon projects into a real-world product—GhostWriter AI Pro. With a few screen taps, GhostWriter Pro AI will provide results that give more refined results without the need to write a complex prompt. 

Here’s what’s behind our generative AI curtain:

• No-prompt Engineering: Learning all the ins and outs of providing prompts to the AI is skipped with GhostWriter Pro AI. Set your tone and output length without the pain of learning prompt engineering. 

• Tailored Writing Styles: The app offers personalized writing styles that adapt to your voice and preferences, providing a more custom and personal writing assistance experience.

• Privacy Features: GhostWriter Pro AI ensures that your data and writing remain confidential. We will never log or train on your data, something many other writing apps do not promise.

• Advanced Editing Tools: Integration of AI-driven grammar, style, and syntax suggestions that go beyond basic spell checking, helping to refine and polish written content effectively.

• Seamless Integration with iOS: Designed specifically for iPhone, the app offers better integration with iOS features, including sharing written content. 

• Customizable Templates: Create custom templates for different writing tasks (e.g., emails, reports, creative writing), which can help you to structure and start new documents quickly.

• Cost Efficiency: Ghostwriter AI Pro is free and supported by ads. The paid version offers a competitive pricing model, providing significant value at an affordable price.

I’d love to know what you think of what we’ve built. You can download GhostWriter AI Pro for free on the Apple App Store.

- Wes

From the BitBakery Blog…

• Developers and users have a love-hate relationship with modal windows. But there is a science behind knowing when to use them. Our Director of User Experience, Attila Schmidt, shares what he tells clients in his latest blog post.

• We recently released a new version of Women’s Brain Health Initiative’s BrainFit – Free Habit Tracker with an integration to Apple HealthKit. BitBaker Annaliese Schropp shares what she learned while working on the project.

• Do you have an ecommerce project in mind? Here is our latest case study on how we helped Folkway Music add ecommerce functionality to their WordPress-powered website.

Ready to take your next project to the next level? We’re here to help.

The idea for the app came to our team during an internal hackathon last year. Writing prompts for generative AI tools like ChatGPT can be complex, and we wanted to create an easier (and more powerful) way to create better content. The hackathon project really got our wheels spinning, so we decided to bring it to the Apple App Store.

You can read about how GhostWriter AI Pro came to be on our blog.

So, how can you win a license? Follow our LinkedIn page and comment on this post by Friday, April 12, 2024, at 11:59 P.M. EST. If you already follow us, great! Just leave a comment, and you’ll be entered.

We’ll randomly select ten (10) winners on Monday, April 15, 2024, and contact you by email. GhostWriter AI Pro is only available for iOS devices at this time. You must reside in a country where OpenAI ChatGPT is available.

- Wes

Eclipses have always held people’s attention. Some view them as omens—both good and bad. 

Today, we’re taking the eclipse as a good sign. April 8 marks ten years in business for BitBakery. Ten years is a major milestone for any business, even more so in technology. Thanks to you, our clients, and our business partners and team members, we've reached this milestone.

This has also allowed us to look back at how our team has grown and evolved to meet changing demands. When Joe, Ryan, Attila, and I decided to start the business, we knew two things: We loved solving problems and liked working together.

Ten years later, those two things are as true today as they were on April 8, 2014.

We also love data and metrics. After ten years, we have:

• Completed over 120 projects.

• Written over 2.7 million lines of code.

• Worked with over 80 technologies, from cloud computing to artificial intelligence to weather APIs.

• Donated 118 times with Canadian Blood Services.

• Called five office locations home.

• Baked over 400 loaves of bread (yet we’re still not a bakery).

Thank you for ten years of trust and business—we look forward to working with you for another ten years!

As always, if there’s anything we can do to assist you with your software design and development needs, we’re here to help.

- Wes

From the BitBakery Blog…

• Building software means we do a lot of demos. At one of our recent Knowledge Nibbles events, I took the team through my process to demo like a pro.

• From yoga to free headshots, our local co-working spaces offer more than just a place to put your laptop. Here’s our updated list of co-working spaces in the Waterloo Region.

• Google Chrome has a new challenger in the Arc browser. What made our design lead, Attila Schmidt, switch to the AI-powered tool? He tells all in his latest blog post.

Ready to take your next project to the next level? We’re here to help.