For over 15 years, I’ve built high-performing sales teams and coached individuals at every career stage, closing over $100M in software sales. Now, I help sellers accelerate their growth, leaders build resilient teams, and founders scale revenue engines through proven GTM processes.

My clients describe me as process-oriented, an excellent communicator, and someone who helps them unlock their potential.

If you’re a seller, leader or founder seeking guidance to accelerate your outcomes, let’s schedule a call via the featured section below!

Mats Uddenfeldt

If you want to break into the top tier of software sales in 2026, you don’t need more theory. You need mindset, skills, and a toolkit that works right now.

That’s why I’ve enjoyed spending time with the EliteGTM Academy over the last few months.

What stood out to me wasn’t hype or polished soundbites.

It was how practical and current the thinking is.

Grounded in how deals are actually getting done today.

I kept catching myself thinking: I could have used this.

Not just as an IC, but during that messy middle phase… when you’re expected to produce, coach, and think strategically at the same time, often without much support.

The Academy makes it easy to establish a stable foundation for your org and team.

What also matters is who you’re learning alongside.

The Academy taps into the wider EliteGTM network and hundreds of operators who’ve been in the seat and are still in the game.

That shows up in the conversations, the examples, and the way problems are tackled.

Less theory. More signal.

If you’re serious about sharpening your craft, this is the kind of environment that actually moves the needle.

Happy to share a sneak peek if you’re curious.

Driving GTM Excellence for Sellers, Leaders & Founders.

Original daily inspiration: Tech Leadership | SRE | Growth

See https://www.alexewerlof.com/who or Google my name.

Alex Ewerlöf

The first team that solves these 5 problems has either reached AGI or is too close:

1. AI that reliably identifies knowledge gaps
2. AI that asks the right questions
3. AI that can reliably validate the answer and go to step 1 as needed
4. It solves this [essentially] search problem efficiently
5. AI that remembers what it has learned and evolves but also can unlearn in face of new contradictions.

There are ways to mimic that with CoT (chain of thought) and symbolic AI (eg. I paired LLM to Prolog with some success) but it fundamentally needs a different loop than next token prediction to succeed in reasonable time and an acceptable price point.

PS. No AI was used in this post. Artwork by my son (he's curious to see how many likes he can get 😄)

💎Follow for original thoughts on Reliability Engineering, Technical Leadership and Growth

"Never argue with a fool" is more relevant now than ever. As the quality of AI improves, more and more people delegate their thinking to it.

Brain is like a muscle: the more it's used, the better it gets.

Some time ago, I had a conversation about a document with a friend who takes pride in riding the AI wave.

🫣 The whole thing was slop and screamed AI generated and low effort.

I was asked to give feedback so I did my best but soon I realize that I'm being used as the HIL (human in the loop). The friend had outsourced his thinking to AI, and now the review process was outsourced to me.

I felt sad and used but I decided to ignore the feeling and invite an intellectual discourse with some basic questions about the AI generated ideas.

He took it personal very quickly and went to defence. All of a sudden I was "criticizing" him not the work.

Then it occurred to me: the AI sycophancy has lowered the tolerance of a civil debate. We had it comong: after years of social media curation and echo chambers now AI has pushed it to the next level.

"You're absolutely right", I wanted to reply and end the feedback session. But I saw he had already left. 🤣

So what did I learn:

1. The energy I put into giving feedback shouldn't exceed the energy that went into creating the target artifact.

2. Be mindful of AI brain rot and sycophancy at scale. Do your best not to fall in the same hole but also be equipped to deal with it because you can't avoid the majority.

3. As someone who builds AI products and teaches it to others, I feel responsible to highlight the risks but also find a way to reduce the negative impacts of AI overdose.

Do you have any ideas for the last one? Putting a disclaimer cigarette-style isn't it. I still find AI useful but it's a double edged sword. How can AI engineers reduce the negative effects of AI overdose on society?

Slapping "engineering" on an activity doesn't turn it to an actual engineering discipline!

The so-called "prompt engineering" is a set of techniques to work around the limitations of language as a communication medium and LLMs as a tool to process language.

It can be mastered in a few hours reading a few blog posts, cheat-sheets, watching some videos and lots of trial and error.

It's a useful skill.

But calling it engineering is far fetched. 😆

Prompt engineering is as "engineering" as playing music on a music player is considered being a musician.

🎯💯 This is gold! 😄 I see too many junior/mid level engineers tearing themselves apart to get their foot through the doors of big tech. ex-Big-Tech influencers made a career out of it. Well, my friend, I don't know your exact life situation but don't be in the business of selling your time (a finite resource) to the highest bidder. Many of these companies have nasty culture (far beyond what's in this post) and mistreat the customers and their society, do you REALLY think they treat their employees any better? What's the alternative? Put your finite time helping companies that solve an actual problem. They may not be super-se*y or pay like a corporation that dodges tax and manipulates national and international politics, but money isn't everything. What's the price of an entire human life, wholesale?

Director of Analytics & BI | Customer Experience & Support Analytics | Decision Intelligence | Neurodivergent Systems Thinker | Continuously Curious

Hopefully something to lighten the weight, Top Things EX-Amazonians Will Not Miss

-Writing six pagers that are never read past the first paragraph.

-Wondering to yourself why is it called a six pager when it has 38 pages of supporting data.

-Having to write a two pager upon request to explain your six pager.

-Meetings about meetings that could have been an email, followed by an email scheduling another meeting to talk about scheduling next meetings.

-Calendar Tetris that somehow still results in back to back meetings from 7am to 6pm.

-The phrase “Let’s take this offline” meaning “Snap, This is about to get political.”

-Chime notifications triggering an immediate stress response, even on weekends.

-Being told to “Bias for Action” right after being punished for acting.

-Metrics that matter deeply until the org changes and they suddenly do not.

-The quiet fear that you optimized the wrong KPI and will find out three quarters later.

-The mental gymnastics required to explain why headcount is frozen while workload is not.

-Pre-reading documents longer than most short novels.

-Spending more time aligning stakeholders than solving the actual problem.

-The phrase “We need to socialize this” killing all forward momentum.

-Pizza being referenced constantly but rarely, if ever, actually appearing.

-Knowing the org chart better than the product.

-Being told to “disagree and commit” after the decision was already clearly made and finalized.

-Playing AtoZ Forensics Detective to figure out root causes of persistent problems.

-Having a title acronym like BIE and not being able to articulate how what you do is completely different than the next 10 BIEs.

-Feeling like you have to earn the right to think deeply.

-Explaining to non Amazonians what a six pager is and watching their face fall.

Add your own in comments, I am absolutely certain there are many many more......

If you are reading this and today was your last badge swipe, breathe. You are not broken. You were trained in a very specific system. There is a lot of life and work beyond it.

If you want to laugh, vent, or compare notes with someone who has already crossed the bridge, I am around.

#Amazon #AmazonBash

🤣🤣🤣 This is a recursive image! The "author" put his own name on AI generated slop acknowledging the value of giving credit while using tools that are trained on illegally obtained work of art. It's most likely generated in DALL-E (by OpenAI which already has several legal cases and class actions open against them by artists, journalists, authors, and content creators).

Hammer, you say. 😄 Instant block from my side.

Industry 4.0 & Digital Transformation Enthusiast | Business Strategist | Avid Storyteller | Tech Geek | Public Speaker

“Hey guys, I have the solution to all of our problems.”
That sentence is doing a lot of work in 2026.

Right now, AI is the biggest hammer most organizations have ever picked up. And the moment that happens, everything in the room starts to look suspiciously like a nail.

We are confusing a 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐭𝐨𝐨𝐥 for a 𝐮𝐧𝐢𝐯𝐞𝐫𝐬𝐚𝐥 𝐨𝐧𝐞.

I’m seeing this everywhere. The conversation shifts from “What are we trying to do?” to “How fast can we roll this out?”  And suddenly, the hardest parts of the job seem to disappear (judgment, prioritization, trade-offs, etc.).

AI has a way of doing that. Not because it replaces thinking, but because it creates the illusion that thinking is done.

AI is exceptional at certain kinds of work:
• Pattern detection across large volumes of data
• Summarizing, generating, and translating information
• Supporting human decision-making when the context is clear

But it’s terrible at others:
• Resolving unclear ownership
• Fixing broken processes
• Making trade-offs when incentives are misaligned
• Creating accountability where none exists

Those aren’t AI problems. Those are organizational problems. And intelligence doesn't fix a confusion and alignment problem.

This is why some teams “deploy AI” and see very little impact. The tool works. The environment doesn’t. So the hammer swings… and the screws strip, the anchors crack, and everyone wonders why the wall looks worse than before.

The companies getting real value from AI aren’t using it everywhere. They’re using it precisely. They’re matching the tool to the problem, not forcing the problem to fit the tool.

𝐈𝐧𝐭𝐞𝐫𝐞𝐬𝐭𝐞𝐝 𝐢𝐧 𝐝𝐢𝐯𝐢𝐧𝐠 𝐝𝐞𝐞𝐩𝐞𝐫?  https://lnkd.in/e7dTek4f

*******************************************
• Visit www.jeffwinterinsights.com for access to all my content and to stay current on Industry 4.0 and other cool tech trends
• Ring the 🔔 for notifications!

✊☁️ "DON'T HALLUCINATE" 👈 this is garbage.

🔴 You may have seen people recommending to add "no hallucination" to your prompt but this is garbage advice! If only things were that simple!

You have to understand how LLMs work! They don't intentionally hallucinate, and despite being trained, verified (e.g. RLHF) and tested (Evals), they still do that.

The best way to think about the current generation of AI is what happens to your brain when you're driving mindlessly or when you experience lucid dreams. These things are definitely not "alive" and "aware" in the human sense. They're just a bunch of emulated neurons modeled after [what we know about] our own brains. They do system 2 thinking (there are ways to pair them up with code to add system 1 thinking and I've listed 30 such patterns here: https://lnkd.in/dcxa_b64)

🟠 Asking AI not to hallucinate is as effective of asking your cat not to scratch the sofa! You can put it in all caps to the same effect! 😼

🟢 So what's the solution?

Most raw LLMs do a one-pass in their token-emission stage. Essentially they don't go back and reflect. The emitted tokens are the parts you see, internally, they arrange their tokens ("thought") before inference.

You can use that fact to make them smarter. Below is the ORCA framework. It looks like code (and is actually written by a LLM to guide other LLMs). This is just one solution but I've been using it for a couple of months and can vouch for it.

The answer will be longer (so it'll cost more) but by explicitly forcing the agent to think through, you can increase the quality of your one-shot prompts. Plus, you get a peek into the AI's reasoning and correct it as needed.

There are many more techniques, but this is just a base line. Go to that link, copy the instructions and paste it to your Gemini Gem or Claude/OpenAI Project.

If it works for you, a reshare ♻️, like 👍 or Github star ★ can help others get an idea of its quality.

As usual, this is open source. Steal, tweak, reshare, or sell. No credit is required (although it'd be nice). There's more where that came from. 💪

https://lnkd.in/davWp7gq

Just a regular reminder that I'm against the "pay to play" model of LinkedIn and will NOT pay a dime for a paid profile to shove my content to people's eyes.

👍♻️ If the content is good, it'll spread on its own. Speaking of which, I fully rely on your likes and reposts to share it if you find it useful.

Also, if you don't want the "algorithm" to decide what meets your eye, you can go to anyone's profile and choose "All" from the bell notification 🔔

LinkedIn charging for exposure is the NUMBER ONE reason there's so much garbage on this platform. I'm not desperate enough to pay for exposure, sorry LinkedIn. 😄

taplio

The Battle

Mats Uddenfeldt

Alex Ewerlöf

Followers

Posts

Engagements

Likes

Performance ScoreMats Uddenfeldt's LinkedIn Performance

Performance ScoreAlex Ewerlöf's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Mats Uddenfeldt

Alex Ewerlöf

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Mats Uddenfeldt's LinkedIn Performance

Performance ScoreAlex Ewerlöf's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreMats Uddenfeldt's LinkedIn Performance