Excited to share the 2026 Words of the Year, AI edition:
1. Substrate
2. Scaffolding
3. Unpack
4. Honestly
5. Sharp
📰 BREAKING 📰: @ScuderiaFerrari lodge appeal in relation to @LewisHamilton’s podium finish in Malaysia’s Bahrain Grand Prix
Be very interesting to see how this plays out 🍿
#f1 #FormulaOne #ferrari #hamilton44
🤖 Made with AI
Tim Norris-Wiles retweeted
🔥 Our hype reel for @G2Eshows 💰
One thing about this space. The amount of data and analytics related to marketing is profound. Every transaction and every behavior is captured. This is exactly why we are having so much success here.
Add columns to a table and LLM predictions start to fall apart.
@LTMpredict found they drop to majority-guess levels by ~30 columns, while classic methods improve.
Customer data is FAR wider than this.
Research on what this means for marketing coming soon from @neuraliftoff
Have to say it was a big surprise to see Mr Tumble on #granddesigns this evening. And that he has a wife!
My Opus 5.5 initial thoughts:
1) It's quick. Feels a bit like 4.6 did which is high praise.
2) Less verbose in responses. Finally.
3) Prose generation is a massive improvement. Opus 5 was AWFUL at writing and it feels like this fixes it.
4) Efficiency has gone through the roof.
Interesting to see news of Harvey's margin collapse. On surface an incredible product + team, + getting to 15bn value is laudable in a number of ways. Just surprised to read that lawyers don't like using it + their UX sucks... I assumed UX / harness would be their moat 🤷♂️
Find out why the addition of a neural-net layer to your customer data stack is becoming the norm for performant brands. Download our white paper with SAS here.
neuralift.ai/white-papers/wa…
Tim Norris-Wiles retweeted
Too many AI personal assistants, too little time to assess all of them. Nice work here: assistantbenchmark.com
Tim Norris-Wiles retweeted
We've collaborated with @SASsoftware Customer Intelligence 360!
Discovery>Decision>Action is a closed-loop framework for Marketing AI that begins at the data warehouse, touches the customer in real-time and gets smarter over time.
Interested in SOA Marketing AI? Read this 👇
Imitation is the most sincere form of flattery, they say.
Nice to see Apple jump on the clay bandwagon with @neuraliftoff this week! 😏
Huge. Nemotron is going to get serious distribution control out of this, and it feels like the timing for N4 probably key.
techcrunch.com/2026/08/26/nv…
There's upside hiding in your data, it's just hard for marketers to find without help!
Using the same old tactics won't deliver anything new, and waiting for the data team to rescore the base means you're in Q4 already 😫
Thankfully there's another way: neuralift.ai
Get a lot of people asking me to explain "Lift Potential" mode in @neuraliftoff, so here goes.
Lift Potential works with transactional data in your warehouse to identify where to prioritise... It changes our segment discovery view to be ranked by $£€ upside, not population...
This matters b/c historically the way we approached marketing was to spend money first, read the results, and then optimize towards the best performing group or groups.
Lift Potential allows you to see where the likely growth lives BEFORE you have spent a penny 👀
You can even game out upside scenarios using the built-in modeller, which allows you to assess worst / good / best cases, with everything is based on your 1P data; no presuppositions, biases or guesswork.
Combines beautifully w/ our new Actions feature! xitter.cf/jonathanmendez/status/…
We've been working all summer on turning customer 360 tables in @Snowflake @databricks @GoogleCloudTech and CDPs into "Ranked Actions."
This builds off the outputs of our Neural Network Segmentation we do with @nvidia GPUs.
More to do but we are getting close to release.
Just one of many releases and upgrades we’ve been working on at Neuralift.
The product is becoming quite uniquely powerful, in moving from insights into actions!
We've been working all summer on turning customer 360 tables in @Snowflake @databricks @GoogleCloudTech and CDPs into "Ranked Actions."
This builds off the outputs of our Neural Network Segmentation we do with @nvidia GPUs.
More to do but we are getting close to release.