Cloudly

https://www.washingtonexaminer.com/op-eds/4638304/larry-sang...

I've had success using the underlying harness - pi-mono as a data analyst in a sandbox.

Ever since the control bug bit me in my EE undergrad years I am happy to see how useful the knowledge remains. Of course the underlying math of optimization remains general but the direct applications of control theory made it much more appetizing for me to struggle through.

Well publicly is a question of where. There's still a lot of startup talk on startup focused reddits / twitter / hn. The main trend I have seen with AI tooling is the dialogue shifting to indie / small teams rather than the VC rush of the 2010s.

https://x.com/bcherny/status/2007892431031988385?s=20 Seems to be moved to the default now. PSA for anyone who didn't see

There Is No Antimemetics Division - qntm

Just finished For Blood and Money - Nathan Vardi. Enjoyed the real world characters who drove the developments of cancer research forward and the messy business of pharma.

I enjoyed Ra - more than "There is not anti-mimetics division" which I felt lost itself in the second half.

I greatly enjoyed Anathem - though I have always been a sucker for Stephenson.

I found 2666 to be profound and tragic. Obviously the intention of the book - but given that the crime rate in my country is analogous to Mexico's - the pain of the femicide in the book and casual cover-ups felt tangible to me. I feel it is clear that is is incomplete due to the Author's death, but I don't think that takes away from the book.

In a similar vein, but in completely different genres I finally finished the wheel of time series which I started nearly 15 years ago. After getting to book 11 I pushed through the rest this year. Unfortunately I felt mostly underwhelmed. I think the Sanderson transition did not age as well as initially received. Maybe it's because I have read too much Sanderson before getting to him. His voice is noticeably different to Jordan's.

Finally I really enjoyed both the Wager and Pachinko.

I have started a slew of other books and not completed them but there are pending standouts. I am still halfway through Stalin by Monetfiore, and I enjoy returning when I can stomach the horrible history.

Honorable mention to Apple in China which I did not complete and might eventually. Don't think I will finish Fear and Loathing in Las Vegas though.

In the current growing pending pile I have 1493 - which I know I will enjoy, everything is tuberculosis and now it can be told. We shall see if I make any headway :)

I have been through a huge over supply of AI generated CVs using similar tools this year. I am sure this will help people so not bashing the tool per se, but bare in mind that you will be joining the robots.

I don't know what the solution really is, but as much as I hate it, a verified Linkedin Profile is a huge signal now. Otherwise a real personal site with care and history. I am guilty of being bad with the second so - no easy answers. Maybe even a wacky looking CV would make me notice a candidate more. Powerpoint 90s style.

The recent novel Abundance seems to be agressibley grouped with the John Green novel An Abundance of Katherines - which I think is a humorous retelling of 2025 but also maybe needs some matching work

Metly | Founding Engineer / AI Engineer | Remote (UTC ±2) | https://www.meetmetly.com/

Metly is a pre-seed startup (team of 5) building strategic intelligence for pharmaceutical companies. We're aiming to replace consultants. Pharma companies hold critical knowledge across disconnected systems—regulatory filings, clinical pipelines, competitive intelligence—none of it linked to real market activity. We're building the layer that connects it all and makes it actionable.

You'll be working on the hard parts of getting LLMs to construct knowledge graphs users can actually trust: entity extraction from dense regulatory documents, grounding outputs with verifiable citations, and turning unstructured data into clear recommendations.

Stack: TypeScript/Python, React on TanStack, PostgreSQL, Kubernetes on GCP. We have traction and are moving fast in 2026. Competitive salary and meaningful early-stage equity.

graeme(at)meetmetly.com

I don't get the hype for a specific agent here really - most cloud sdks _+ pulumi support languages like typescript and python for IaC. No YAML. I think terrform was thinking about this the last time I was working with them. That gives you all the benefits of software agents we already have available. I think for most those will be enough.

I have found that a useful method for getting distractions off my work laptop has been https://selfcontrolapp.com/ (if you're on mac).

Mathematica: A Secret World of Intuition and Curiosity- David Bessis

https://www.goodreads.com/book/show/200128457-mathematica?ac...

Excellent book on mathematical thinking in the true sense - what needs to happen in the mind's eye to really grapple with abstract mathematics. Definitely a eye (mind?) opener for someone who has some graduate level math education but couldn't gel with the crazier stuff.

Came across the book from this article which was on HN a little bit ago: https://www.quantamagazine.org/mathematical-thinking-isnt-wh...

https://www.unspun.io/ seems to be trying this

I would recommend "The Things We Make" for an outlook into engineering mindsets through history. A good reminder that a lot of useful engineering comes before the theory can fully explain it :)

https://www.goodreads.com/en/book/show/75598048

https://github.com/spcl/graph-of-thoughts Code for the paper too

This mostly makes me worry about how it would be misused in my country. But I suppose if you have enough of a high trust society it could be fun

The new movie you're thinking of is based on Killers of the Flower Moon: The Osage Murders and the Birth of the FBI by David Grann

I'm free of this pain currently but I could see this saving a huge time pain in the product -> engineering handover step. Very cleanly done. Honestly great just for making sure engineering designs have covered all paths.

There's a measure for that but I can't remember how's it called Slew rate for opamps I believe

Also worth noting that the "recursion" without negative feedback will hit a limit when the op-amp is saturated. Similar to a "stack overflow" I guess. Things mostly won't break but you'll get a flat output at the positive or negative rails of the op-amp - which causes some of the pain of analog signal processing.

Semantic search using CLIP embeddings / other LLM embeddings could be an amazing addition too.

Would also recommend paperspace gradient if you're wanting decent access to a GPU at $8/month for a pro-sub. Been great to help with learning before diving head first into buying one.

An interesting theory as to why spatiality helps so much is outlined in Jeff Hawkins' "1000 brain theory" (https://www.amazon.com/Thousand-Brains-New-Theory-Intelligen...). His team's research hints that all neocortical processing happens with reference frames at the core.