@pkozlowski.bsky.social
Open source, Angular.
Learn how Google Security Engineers think about securing web apps and how Angular has built-in features to help protect against attacks in this new blog: bughunters.google.com/blog/5457130...
who would think, huh?
;-)
Spent the last 4 months looking into LLM codegen quality and techniques of improving it for the new APIs (not in the training set).
Many learnings from the process, with the key one being: LLMs "preffer" good API design.
Codegen fails? Your human devs would be likelly confused as well!
Would love to collaborate on plugging RippleJS into github.com/angular/web-...
We did quite a bit of Angular syntax fine-tuning + using the tool to drive design of some new DSLs (with some the ideas very similar to RippleJS).
@devagr.bsky.social was also toying with it for SolidJS.
Interested?
Time to go back home after two weeks of travel and speaking about @angular.dev .
Thnx for having me
@angularconnect.com and WhatTheStack conference!
Got a couple more talks lined up this year but that's a story for another post.
A good opportunity of writing a series of follow up articles, the whole "Perils of alternative X" saga :-)
More seriously, the article does a great job of pointing out some friction points with signals. But IMO those are easier to live with when compared to friction points of alternatives.
It is a popular opinion, at least in my circles :-)
My main reasoning is that library-agnostic widgets need to compromise to fit into different frameworks and not making full advantage of the ecosystem.
Plus "generic" design is often harder, makes lib iterations slower etc.
I'm so happy that @trueadm.dev shares his learnings in the open. There are 2 problems very close to my hart being explored here:
- reactivity / rendering dance;
- modern authoring format.
Some will claim that UI fwks are "solved problem". I don't agree. There is still so much room for improvement!
Putting finishing touches on my "Angular: framework of the future" talk (wts.sh/sessions/8iw...) for
@what_the_stack
conference.
Arriving a few days earlier to explore Skopje was a great move - loving it here!
Tune to this one if you've got a chance. We will be sharing some LLM-related tools and techniques that I'm particularly interested in atm.
15.09.2025 16:34 β π 1 π 0 π¬ 0 π 0
We did lots of experiments with LLMs generating
@angular.dev
code. They are _great_.
Totally should be part of the #Angular community, proudly wearing a #Angular T-shirt.
In general:
- APIs that are confusing for humans tend to be confusing for LLMs;
- APIs that are not flexible enough are more prone to hallucinations;
There are so many interesting techniques for using LLMs as user-study participants of sorts. A true LLM-driven design workflow.
LLMs are _really_ good at writing code - not a news at this point.
What I find more interesting is LLMs used as beta-testers for new APIs and DSL:
- start with idea + docs + examples
- LLM-generate lots of code
- see what and how it fails.
It seems like LLMs tend to fail _similarly_ to humans
1/2
I've got so many great memories from the previous editions of #AngularConnect and couldn't be happier to see it happening again.
The hardest part of it all: select keynote material. With so much stuff happening in #Angular we will need to do some ruthless "scope cuts" - a good pb to have I guess?
Yep, the Chrome DevTools Perf panel extension point is _great_. We clearly saw its potential and this is why @angular.dev ships with the tight integration.
More info in developer.chrome.com/blog/perf-pa...
Vercel's new AI SDK 5 introduces support for Angular so you can supercharge AI-driven development with the latest Angular features π
Blog post: vercel.com/blog/ai-sdk-5
Code sample: github.com/vercel/ai/tr...
π’ We're finally looking to ship Observables in Chrome!! It's been a long time comingβmany thanks to all in the community for the great feedback and help along the way!!
Especially @benlesh.bsky.social, @keithamus.social, @domenic.me, @jordan.har.band & more! π
groups.google.com/a/chromium.o...
I'm so happy someone captured Angular's story in a documentary!
This 15+ year tale is summarized in just an hour...
This video shares a glimpse into the project into which we have poured thousands of hours of love and dedication.
Enjoy!
www.youtube.com/watch?v=cRC9...
I would _love_ to read a high-level overview focusing on:
a) your initial motivations (dissatisfaction with the mainstream? something else?)
b) main ideas / initial story;
c) why it got abandoned / what was the insight leading to that decision.
Hmm, this wouldn't be my immediate conclusion.
For one, after refactoring many apps, I see RxJS overused - where a simpler approach (signals, promises etc.) would work great.
Also, the upcoming Observables spec my change how we think about events processing and integration with signals.
Actually, this one of the thing that we very much do NOT want to address with signals - this is because signals are not meant to model those (eventing) situations.
This requires adjustment in thinking but a reactive graph should describe relation between data and not a sequence of events.
I wish this would be a more universal knowledge.
I've profiled hundreds of apps. None of them was anywhere close to the tech limits. Perf of all was improved by relatively simple treatment of removing obvious bottlenecks.
Most of those apps were never profiled until someone complained....
![Angular: The Documentary [OFFICIAL TRAILER]](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:rbdhgb34fptx5zjrsqzcoqch/bafkreihcoli7h6lju5cp6wfasacvds36k6dnf6lsxzyk3pvey3xuue2b2q@jpeg)
There is so much passion, energy and history behind @angular.dev. There are many ways to tell its story and different people will have different stories.
But I can very much identify with the story told in the Angular documentary: www.youtube.com/watch?v=nONH...
Humbled and proud to be part of it!
I own large part of my IT knowledge to the books published by Manning.
But is is very special, and I must say - somewhat surreal - to see a book that talks about concepts and APIs that I've co-designed.
"Modern Angular" by @armandotrue.bsky.social
www.manning.com/books/modern...
Yay, hi @angular.dev !
22.01.2025 19:35 β π 4 π 0 π¬ 0 π 0This is all generalisation of course and libs using "block VDOM" approach have a significant advantage as those recognise parts of UI that will _never_ change.
09.01.2025 20:40 β π 1 π 0 π¬ 1 π 0This is a very good take on it. Agree that there might be no clear answer.
But one thing is clear: VDOM scales with the overall "size of the UI" (you need to diff since you don't know what changed) while signals scale with the number of changes.
Perf will depend on the type of UI being built.
Wake up friends, we just dropped a new video! π₯
If you're using Observables and Angular, you'll love this discussion about the future of Observables featuring @domfarolino.com (Chrome Team) plus @synalx.bsky.social, & @pkozlowski.bsky.social (Angular Team). #WeLearnTogether π€
youtu.be/ZXnwj1B29oc
Not a React answer but the principle is the same: I like to think about data fetching as a mapping of a reactive state => reactive response.
Newly introduced Angular's resource, SolidJS resource (minimal APIs) and TanStack Query (a real lib) lets one think in the "mapping" mental model.
Ideally we _hope_ to find a small, cohesive and composable abstractions. But this is hard in practice: requires deep understanding of typical and corner use-cases
It is a slow and iterative process. And this is why we will see both new and substantially changing frameworks in 2025.
3/3
