Tim Hanlon

1.4K posts

Tim Hanlon banner
Tim Hanlon

Tim Hanlon

@timhnln

Building a better CMS @useMinima

Melbourne Katılım Ocak 2009
1.3K Takip Edilen1.3K Takipçiler
Tim Hanlon
Tim Hanlon@timhnln·
Getting your clanker to clone repos as references works really well, but I've been playing around to see if there's a better/faster way. This Effect v4 code search tool started as an attempt to find faster answers to "how do the people who know what they're doing do [thing]", and ended up being useful for me as well. It's end-to-end Effect too - Alchemy v2 for Cloudflare infra & Foldkit for the UI. A bit more deslop/polish and I'll ship it.
Tim Hanlon tweet media
English
0
1
19
2.9K
Mint Blitz
Mint Blitz@MintBlitz·
Marathon players will see this and be like hell yeah
English
129
109
2.6K
178.3K
Tim Hanlon
Tim Hanlon@timhnln·
I built a voice agent for the Nintendo 3DS
English
0
1
3
428
Tim Hanlon retweetledi
pythonxi
pythonxi@python_xi·
Meet Dory: a free, open-source Docker Desktop and OrbStack alternative for Mac. One shared VM for all your containers (~4.7x less idle RAM), a real docker socket, one-click Kubernetes, full Linux machines for all dev work. ~6 MB, no Electron, no accounts, GPL-3.0. augani.github.io/dory
GIF
English
48
138
1.2K
90.6K
Tim Hanlon
Tim Hanlon@timhnln·
@_overment This looks really cool. I keep saying that we haven't found the end-game UI for agentic development yet, love seeing people's unique takes on it.
English
0
0
1
93
Adam
Adam@_overment·
the second iteration of my software factory. I even came up with the name, Zion, to highlight the fact that humans still matter in programming. also, I moved toward a UI that shifts even further away from a chat UI and focuses on controllable steps, where I know exactly what is in the context at any given moment. The context window itself stays almost empty most of the time, except for necessary components that matter from a prompt-caching perspective as well. a library of triggers, hooks, agents, and skills is also something I was missing. a quite important thing here is that the system is designed so that the main agent can change pretty much everything in the settings. Meanwhile, I can jump between pretty much every single primitive using the command palette. @kskrygan this is a preview of how I imagine a next-gen IDE that is focused on outcomes and actual craft. I still believe that either you or the folks at xAI can deliver software that lets us actually work together with agentic teams, without giving an illusion of control or flooding us with tons of output that is 95%+ the model yapping, not an actual signal.
Adam tweet mediaAdam tweet mediaAdam tweet mediaAdam tweet media
Adam@_overment

f-ck me. looks like in 24 hours I’ve created a tool that’s made all coding agents obsolete for me. still iterating, but at this point I’m more than excited about the results I’m getting. this tool is based purely on my understanding of how LLMs and agents work, so it’s full of deterministic hooks and checks, but also dynamic triggers that can be picked by either me or the agent. for example: things like “skills” implemented in, let’s say, Claude Code are deeply flawed because they expect the LLM to follow additional instructions in a thread where there’s already a ton of other context. the same goes for system prompts and other “rules” we want an agent to follow. and the funny thing is that I don’t expect the typical user to be able to work with such a tool effectively, since it requires deep knowledge of what’s going on under the hood. it’s obvious to me why big labs or IDEs don’t follow this path YET, but I believe this is the direction we’re all heading as general awareness of language models and agents improves over time.

English
3
1
23
5.5K
Tim Hanlon
Tim Hanlon@timhnln·
Experimenting with compaction in Arc Work today. Big picture goal is the always-on "chief of staff" agent.
Tim Hanlon tweet media
English
1
0
1
118
Wes Bos
Wes Bos@wesbos·
Are you using some sort of app to manage your agents? Herdr, conductor, solo, warp, tmux, superset, emdash... Big space right now, every tool covering different pieces
English
227
13
534
85.1K
Tim Hanlon
Tim Hanlon@timhnln·
Claude spawning another Claude, and assigning it work
Tim Hanlon tweet media
English
0
0
1
107
Tim Hanlon
Tim Hanlon@timhnln·
I'm honestly shocked at Gemma 4 E4B working this well with near zero effort on my part. An 8GB model running locally via Pi and LM Studio. Fast and useful.
Tim Hanlon tweet media
English
0
0
0
133
Tim Hanlon
Tim Hanlon@timhnln·
Happy accident from ingesting Claude history - always hated not being able to see what it was doing with its memory
Tim Hanlon tweet media
English
0
0
0
69
Tim Hanlon
Tim Hanlon@timhnln·
Experimenting with orchestration in Arc Work. I told Claude that relying on me for testing was the clear bottleneck, so it went and built a headless harness it can drive without me. Added Pi support so that we can use local models too.
Tim Hanlon tweet media
English
0
0
1
86
Tim Hanlon
Tim Hanlon@timhnln·
Reworked sidebar shows active sessions across all projects & worktrees
Tim Hanlon tweet media
English
0
2
6
841
Jason
Jason@JasonHassold·
@0x15f custom with AI is better than any library
English
3
0
2
1.9K
Jake Casto
Jake Casto@0x15f·
Whats the latest and greatest for chart libraries? ReCharts is massive, wanna drop it
English
30
1
68
28.2K
dex
dex@dexhorthy·
context engineering docs for agentic engineering - plans, research, etc SHOULD NOT be stored in version control: A good docs management system keeps them: > outside your repo > accesible to agent via FS tools > discoverable by agent (even just maybe via sysprompt append) > persisted / recoverable / archivable > collaborative (shareable, commentable) why keep them outside the core VCS repo? 1) they don't need merge semantics, just linear history is plenty in 99.9% of cases 2) if they are committed that means they can live on branches, get lost when you change branches, you have to remember where they were, etc etc wdyt?
English
110
15
497
79.2K
Tim Hanlon
Tim Hanlon@timhnln·
Landed initial worktree support & GitHub PR awareness
Tim Hanlon tweet media
English
0
0
2
130
Tim Hanlon
Tim Hanlon@timhnln·
@whisvue This is very cool, will send this to friends I can't convince to play the game
English
1
0
1
216
whis
whis@whisvue·
Not done with the navigation yet so next steps before deploy are: - Visual changes to nav and grid - Add new Navigation to mobile - Adding a list/grid view toggle I'll continue with overlay improvements after this byeee
English
12
83
1K
26.6K
Tim Hanlon
Tim Hanlon@timhnln·
It's a local-first graph that preserves the work between work/commits/PRs, with everything traceable back to the agent session. GitHub issues can be part of it, but not a great place to store it.
English
0
0
1
21
Tim Hanlon
Tim Hanlon@timhnln·
This is so much better than Markdown plans
Tim Hanlon tweet media
English
1
0
0
110