I disagree, I think it can amplify learning, but it depends how and what you are asking. I build up screening experiments with follow up hypothesis generation to both further my knowledge as well as expanding the solution space. I can get citations with papers covering similar topics to go deeper.
04.03.2026 17:02
π 2
π 0
π¬ 1
π 0
Smatterings of F#
YouTube video by dotnet
Daily Reminder to try #dotnet and also #fsharp since is part of the official sdk and is an awesome functional first language ! youtu.be/GcEHiY6Vp-8?...
25.02.2026 18:52
π 5
π 1
π¬ 0
π 0
Serde.FS.Json is the first backend in the new Serde.FS ecosystem β fast, reflectionβfree JSON for F#.
Powered by FSharp.SourceDjinn, a small engine for building F# source generators.
github.com/fs-djinn
#fsharp
03.03.2026 15:04
π 13
π 1
π¬ 1
π 0
Did that replace the myriad based ast oriented way you used to use?
03.03.2026 18:47
π 0
π 0
π¬ 1
π 0
Fable Β· Announcing Fable 5 Release Candidate
The wait is over! Fable 5 Release Candidate is officially here.
After a year of hard work from the community, we're bringing you:
β¨ .NET 10 & F# 10 support
π A massive Python update (now powered by a Rust core & PyPI!)
π Cleaner JS/TS Pojo bindings
fable.io/blog/2026/20...
#fsharp #fablecompiler
03.03.2026 17:18
π 29
π 11
π¬ 4
π 0
Because the space is changing fast, or something else?
03.03.2026 16:58
π 0
π 0
π¬ 1
π 0
Tell that to the tech hiring companies. I'm thinking of substantially re-training to be on the other side of the wire e.g ML/AI research etc. It's a tricky time, at least for me it is.
03.03.2026 16:07
π 1
π 0
π¬ 2
π 0
I've seen so much anti AI sentiments that I was starting to feel what I have experienced was a fluke. It's great to read that my experience is mirrored in other domains. It's depressing to feel like your profession has been relatively replaced by a subscription so I can understand subterfuge arguing
03.03.2026 15:39
π 1
π 0
π¬ 1
π 0
Academics Need to Wake Up on AI
Ten theses for folks who haven't noticed the ground shifting under their feet
Sorry, Bluesky, but I have to say it: AI can already do social science research better than most professors with PhDs. And, for the first time in my life, I really have no idea what happens in five years.
Things are changing already, we just need to wake up.
03.03.2026 00:08
π 177
π 31
π¬ 304
π 273
I'm seeing a lot of anti AI sentiment, not AI has taken our jobs and stuff like that, but things like: AI agents don't work or they are no good or don't help. I see the exact opposite and cant help that my career as a programmer is defunct, at least in terms of actually writing code.
03.03.2026 13:52
π 2
π 0
π¬ 0
π 0
I'm using a similar flow except rather than tacking tech dept or issues; I'm ablating ML experiments and then using a more advanced model to structure a hypothesis based on the results.
01.03.2026 18:14
π 2
π 0
π¬ 0
π 0
It not really the future its just getting easier to set the guard rails in place so you don't generate poor code. The agentic workflows just add cronjob type workflow to the process so you get iterative advances.
01.03.2026 18:14
π 2
π 0
π¬ 1
π 0
You know, we used to joke about 'vibe coding.' Well... things have changed. We should call it what it is: Coding with AI.
From @karpathy.bsky.social
It is hard to communicate how much programming has changed due to AI in the last 2 months: not gradually but in this last December.
26.02.2026 02:26
π 72
π 3
π¬ 9
π 1
For me the idea and chaining of thoughts to a solution space is the important part. How that gets done is irrelevant.
26.02.2026 13:14
π 1
π 0
π¬ 0
π 0
Well that feeds into guard rails with mcp servers the same as the IDE has these to tell you a syntax error. The fundamental is an abstract idea having a syntactic language constricts that represents it. The test add parameterizations and limits to the construct.
26.02.2026 13:14
π 1
π 0
π¬ 1
π 0
Brainstorm, plan, implement, test. These are all normal software development layers but if you fashion these into discrete agents with roles you get a self correcting path with guard rails.
26.02.2026 11:57
π 1
π 0
π¬ 0
π 0
I was skeptical a few years ago but I can channel and cross cut new ideas to explore a new design space while only guiding and tweaking code if I have the right structures in place.
26.02.2026 11:57
π 0
π 0
π¬ 3
π 0
Things may change as we get further into the age of not writing much code anymore, where the language doesn't matter. I'm guessing it will get even more stratified though.
26.02.2026 11:21
π 0
π 0
π¬ 1
π 0
I mean it's a fine language and I love to use it, but as an independent software developer it's not a viable.
26.02.2026 11:21
π 0
π 0
π¬ 2
π 0
I think F# is already well into past tense now.
25.02.2026 17:19
π 0
π 0
π¬ 1
π 0
Pretty much 95% of all programming will be done with it but the only question is who will pay for the compute.
25.02.2026 17:17
π 0
π 0
π¬ 1
π 0
Learning Rust. Looks like someone took functional and procedural F# or Haskell and transmuted it with the C-like syntax while using the Malbolge paradigm (making everything as convoluted as humanly possible) into a new language difficult as hell.
25.02.2026 04:05
π 5
π 1
π¬ 3
π 0
I translated to English to read too, interesting post. To be honest it echo's with what I have been doing too. I have never used Cursor but Claude Code is always open.
25.02.2026 12:08
π 0
π 0
π¬ 1
π 0
Almost everything is insanely big in comparison to what it actually does.
17.02.2026 15:42
π 2
π 0
π¬ 0
π 0
How is acrobat reader a 820MB download?
17.02.2026 15:04
π 2
π 0
π¬ 2
π 0
Boosting F# Libraries with Automated Agentic AI | Amplifying F#
This week on Friday @dsyme.bsky.social shows us how agentic workflows can continuously improve #fsharp libraries.
amplifyingfsharp.io/sessions/202...
17.02.2026 13:07
π 16
π 7
π¬ 0
π 0
(Re) Introducing the Clef Language
How the Fidelity Framework Led Us to a New Language
Given how much we've been sharing our work with the #fsharp community, we thought it worth summarizing this pivotal decision to establish a new language. speakez.tech/blog/reintro...
17.02.2026 00:13
π 9
π 3
π¬ 0
π 0