Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­

@benbackx.com

Trying to turn my personal data into a superpower || Dad x2 || Long suffering Spurs fan || Built the Mimoto app

191 followers707 following139 posts

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

BEAR HUNT, the new geopolitical thriller from the creators of Sherlock & Co.,Β  is the No. 1 Drama podcast in 12+ countries!! I am so lucky to be part of such another awesome project by these folks! Search for BEAR HUNT wherever you listen to your podcasts.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Nothing moves. Yet everything is changing.

Underrated aspect of having kids. When they use your YouTube profile to watch some vids, your subsequent recommendations for the next week are epic… worlds biggest container ships βœ…, triceratops vs 3 tyrannosaurus βœ…, paper aeroplane tutorials βœ…. Yes please!!

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

If [design] = [art] + [constraint], then [strategy] = [business design] = [inspiration] + [constraint]. The decisions you make, both the + and - consequences of that, are the constraints. These are often called enabling constraints. Articulate, and embrace them. Here is how:

youtu.be- YouTubeEnjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Just published a new article! πŸ‘‡πŸΎ Your company will stop growing sooner than you think. The "Max MRR" metric predicts revenue plateaus based on churn and new revenue. Many thanks for spreading the love with ❀️ and πŸ”!

longform.asmartbear.comMax MRR: Your growth ceilingYour company will stop growing sooner than you think. The β€œMax MRR” metric predicts revenue plateaus based on churn and new revenue.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

I think "context engineering" is going to stick - unlike "prompt engineering" it has an inferred definition that's much closer to the intended meaning, which is to carefully and skillfully construct the right context to get great results from LLMs simonwillison.net/2025/Jun/27/...

simonwillison.netContext engineeringThe term context engineering has recently started to gain traction as a better alternative to prompt engineering. I like it. I think this one may have sticking power. Here's an …

There’s a lot of discussion around energy usage and general inefficiency of LLMs, but I haven’t seen much commentary yet around model routing/prompt re-writing using local models in order to direct a more concise request to an appropriately sized model. Anyone working on this?

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Great example of generate -> validate loop + error analysis > "the most effective route to improve outcomes was brute force: retry steps until they passed or reached a limit. We give the validation errors ... to the LLM and built a loop runner"

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Just published a new article! πŸ‘‡πŸΎ How to discern which models of companies, strategy, markets, and organizations are wise, and which just look good on a PowerPoint slide. Many thanks for spreading the love with ❀️ and πŸ”!

longform.asmartbear.comAll pretty models are wrong, but some ugly models are usefulIdentifying useful frameworks for companies, strategy, markets, and organizations, instead of those that just look pretty in PowerPoint.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

When planning the next few sprints, what strategic objectives you’re addressing, even while also addressing current obstacles? And don’t say β€œgrowth” because that’s an outcome. Need help? Here’s a framework:

longform.asmartbear.comQuarterly strategic planning using the fairytale structureTraditional fairytale structure fits naturally in our brains, and thus can guide strategic problem-analysis, and a plan that everyone understands.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Search & recsys have historically drawn inspiration from language modeling, such as Word2vec for item embeddings, and GRUs and BERT for sequential models. LLMs are no different and have led to updates in model architecture, training data & paradigms, etc. eugeneyan.com/writing/recs...

eugeneyan.comImproving Recommendation Systems & Search in the Age of LLMsModel architectures, data generation, training paradigms, and unified frameworks inspired by LLMs.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Just republished a classic article! πŸ‘‡πŸΎ Competing against big firms? Their revenue streams are their Achilles heel. Learn how smaller startups have a shot. Many thanks for spreading the love with ❀️ and πŸ”!

longform.asmartbear.comWhy it’s nice to compete against a large, profitable companyCompeting against big firms? Their revenue streams are their Achilles heel. Learn how smaller startups have a shot.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

The fall of a tyrant must be celebrated. And while Syria faces great uncertainty, a nation exhausted by war could yet choose the long road towards peace. Here is how econ.st/4feXK1w

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

The aftermath of authoritarian leaders often leads to complex outcomes. The situation in Syria highlights how unpredictable such scenarios can be. @missiongrey.bsky.social provides the tools to analyze these impacts. #MissionGrey #Geopolitics #Geoeconomics #Syria #RiskManagement #BusinessScenarios

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

"In the cozy era, our screens and the related accoutrements of digital life fulfill all of our emotional and sensory needs." Rereading my column on cozy technology; I'll be on CBC at 8:30 AM ET discussing it www.newyorker.com/culture/infi...

newyorker.comThe Fantasy of Cozy TechFrom the β€œcozy gaming” trend to a new generation of A.I. companions, our devices are trying to swathe us in a digital and physical cocoon.

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Really cool website that showcases references on how to use swift's various formatting tools for numbers, dates etc. to achieve very specific localized string formats πŸ”— fuckingformatstyle.com

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Ben Backx πŸ‡¬πŸ‡§πŸ‡ΉπŸ‡­ reposted

Someone Made a Dataset of One Million Bluesky Posts for 'Machine Learning Research' A Hugging Face employee made a huge dataset of Bluesky posts, and it’s already very popular. Each post contains text content, metadata, and information about media attachments and reply relationships.

404media.coSomeone Made a Dataset of One Million Bluesky Posts for 'Machine Learning Research'A Hugging Face employee made a huge dataset of Bluesky posts, and it’s already very popular.