← Back to Blog

This Week in AI: High-Stakes Updates and Why Being Wrong is a Business Strategy

A look at the latest AI releases from OpenAI, Google, and Apple, and why admitting what we don't know might be the only way to build trust.

By · · blog

This Week in AI: High-Stakes Updates and Why Being Wrong is a Business Strategy

The best deal I ever closed started with me saying "I don't know."

It was a cold Tuesday, and a potential client was grilling me on how a specific AI workflow would handle their edge cases. They wanted a guarantee. They wanted certainty in a field that changes every time a developer in San Francisco misses a night of sleep.

I could have bluffed. I could have used a few buzzwords and polished my way through it. Instead, I told them I wasn't entirely sure yet, but I knew how we’d find out.

The room went quiet. I thought I’d lost it. Then the CEO leaned in and said, "Thank God. You're the first person who hasn't lied to me today."

We’re currently drowning in a sea of "unprecedented" updates and "significant" features. If you feel like you're falling behind because you haven't mastered every new tool that dropped this Tuesday, join the club. The real skill in 2024 isn't knowing every tool; it's being honest about which ones actually do the job.

OpenAI o1 and the Death of the Instant Answer

OpenAI recently released their "o1" series, formerly known as Strawberry. This isn't just another incremental bump in speed. In fact, it's slower. It actually "thinks" before it speaks.

For a long time, we’ve been trained to expect AI to be a glorified autocomplete. You hit enter, and it vomits out text immediately. With o1, you see a little thought trace. It’s reasoning. It’s checking its own work.

In a business context, this is a massive shift. We finally have a model that is significantly better at complex logic, maths, and coding. If you’ve tried to use AI for high-level strategic planning or complex spreadsheet logic before and found it... well, a bit dim... o1 is designed to fix that.

The lesson for us humans? If the AI is slowing down to get it right, maybe we should too. Rushing to implement a tool just because it’s there is a recipe for a very expensive mess. If you're looking for more grounded takes on Tech, more articles on AI might help clear the fog.

Google Gemini 1.5 Flash and the Race to Zero

Google has been quietly (or as quietly as a trillion-dollar company can) slashing prices and increasing speed on their Gemini 1.5 Flash models. They’ve drastically reduced the cost for developers.

Why does this matter to you if you aren't a coder? Because it means the cost of "intelligence" is plummeting.

The features that used to be a premium upsell are now becoming the baseline. If your business model relies on charging a premium just for providing access to AI, your margins are about to get a haircut.

The value is no longer in the access. The value is in the integration. It’s about how you hook these models into your messy, human processes to actually save time. As a bit of a Monday treat, you can see how we handle these integrations over at SalesM8.

Apple Intelligence and the Normalisation of the Strange

Apple is finally rolling out its AI features to the public. They aren't trying to build a god in a box; they’re trying to help you rewrite a grumpy email and remove your ex from the background of a holiday photo.

It’s "boring" AI. And boring AI is where the money is.

While the tech world obsesses over AGI (Artificial General Intelligence), your clients just want to know how to get through their inbox without an existential crisis. Apple is making AI approachable. They are taking the "weird" out of it.

If you’re struggling to sell AI services, stop talking about LLMs and neural weights. Talk about the "Reduce Interruptions" feature on the new iPhone. Use it as an analogy. People don’t want a paradigm shift; they want to get to the pub twenty minutes earlier on a Friday.

Why the "I Don't Know" Strategy Outperforms the Guru

I’ve spent the last few years watching people position themselves as "AI Experts." Most of them are just very fast readers of Twitter threads.

There’s a massive vulnerability in admitting that this stuff is complicated. But that vulnerability is your greatest sales tool. When you are honest about the limitations of these tools, your "yes" carries ten times the weight.

Here is what I’ve noticed about the most successful AI implementations lately:

  • They solve one boring problem exceptionally well.
  • They don't try to replace a human; they replace a chore.
  • The person leading the project is willing to say, "Let's test this, because it might break."

If you’re feeling overwhelmed by the pace of change, take heart. Most of the people shouting the loudest are just as confused as you are. They’re just better at hiding it.

Practical Steps for Sane AI Implementation

So, what do we actually do with all this? How do we turn the noise into something that pays the bills?

  1. Ignore the 90%: Most "new feature" announcements won't change your life. Stick to the tools that have already proven they save you at least an hour a week.
  2. Focus on "Reasoning" over "Regurgitation": Use models like OpenAI's o1 for logic and strategy, and cheaper models like Gemini Flash for quick, repetitive tasks.
  3. Be the Adult in the Room: When a client asks if AI can do something impossible, tell them no. They will trust you forever.
  4. Audit Your Chores: Don't look for "AI opportunities." Look for things you hate doing. If you hate it, there’s a 90% chance a model can do it now.

The goal isn't to be a tech genius. The goal is to be a helpful human who uses smart tools. If you’re trying to figure out where your business fits in all of this and want a conversation that doesn't involve a pitch deck, you can book a consultation and we can figure it out together.

Business is hard enough without pretending we have all the answers. The tools are getting smarter, but they still don't have a sense of humour, and they definitely don't know how to admit when they've messed up.

That’s still our job. And honestly? I’m quite happy about that.

---

I cover the latest AI releases and tools every Friday at steventann.com. If you found this useful, there's plenty more where this came from.

---

About the Author

Steven Tann is an AI consultant, author of "You're Selling AI Wrong", and founder of SalesM8. He writes about AI, sales, and running a business from a narrowboat on the English canals. Connect with him at steventann.com.

Tags: AI Tools, OpenAI, Business Strategy, Implementation, Generative AI