Traun Leyden (@tleydn) profile picture

Indie Hacker. Ex-Databricks ML infra eng, ex-Couchbase eng.

Munich, Germany
Joined December 2009
Ok this is nuts
Introducing Magenta RealTime 2, a new open model musicians can play as an instrument! Run low-latency, live music synthesis natively on your MacBook using MIDI, text, and audio. 🎶 We love seeing Google’s open model ecosystem grow!
1
45
The new @OpenAI gpt-realtime-translate model is absolutely mind-blowing!! After playing w/ the API extensively, I can assure you that this demo was not faked or embellished in any way. youtube.com/watch?v=JOu8v6CB…
84
Looks super cool, can't wait to try it! @lennysan Are you planning to add an API or MCP server?
This is a the future "Littlebird is a desktop app that remembers everything you’ve been working on – meetings, messages, docs, browsing - and helps you stay focused, prioritize, recall, and move projects forward. It uses screen reading to understand all the text on screen, for all applications, and uses that context to build a rich understanding of your life: who matters to you, what you're working on, and what you care about this week and this year."
75
Traun Leyden retweeted
🚨: A petri dish of human brain cells just learned to play DOOM
1,697
5,900
4,931
48,089
32,433,611
I mean, what could go wrong?
now my clawdbot lives in my ray-ban meta glasses so i can just buy whatever i’m looking at
48
Sounds amazing for 7B!
NVIDIA just removed one of the biggest friction points in Voice AI. PersonaPlex-7B is an open-source, full-duplex conversational model. Free, open source (MIT), with open model weights on @huggingface 🤗 Links to repo and weights in 🧵↓ The traditional ASR → LLM → TTS pipeline forces rigid turn-taking. It’s efficient, but it never feels natural. PersonaPlex-7B changes that. This @nvidia model can listen and speak at the same time. It runs directly on continuous audio tokens with a dual-stream transformer, generating text and audio in parallel instead of passing control between components. That unlocks: → instant back-channel responses → interruptions that feel human → real conversational rhythm Persona control is fully zero-shot! If you’re building low-latency assistants or support agents, this is a big step forward 🔥
83
Can I use @Sourcegraph Amp in the Zed editor? @sqs
1
1
714
There's never been a better time to be an Erlang programmer
27
ChatGPT kind of has a built-in churn mechanism. They cut me off from using too much Deep Research this week, so now I'm checking out the Gemini Deep Research ..
41
I've integrated @gethuxe into my daily biking commute: 1) Queue Hacker News auto-generated podcast 2) Listen and ride, and enjoy the amazing engaging dialog I think they use Gemini for the voices. It's very easy to listen to, and great content summary, including comments
1
2
628
Another data point to my hypothesis - anything that involves copy/paste between LLMs is basically a product that needs to be created.
As a fun Saturday vibe code project and following up on this tweet earlier, I hacked up an **llm-council** web app. It looks exactly like ChatGPT except each user query is 1) dispatched to multiple models on your council using OpenRouter, e.g. currently: "openai/gpt-5.1", "google/gemini-3-pro-preview", "anthropic/claude-sonnet-4.5", "x-ai/grok-4", Then 2) all models get to see each other's (anonymized) responses and they review and rank them, and then 3) a "Chairman LLM" gets all of that as context and produces the final response. It's interesting to see the results from multiple models side by side on the same query, and even more amusingly, to read through their evaluation and ranking of each other's responses. Quite often, the models are surprisingly willing to select another LLM's response as superior to their own, making this an interesting model evaluation strategy more generally. For example, reading book chapters together with my LLM Council today, the models consistently praise GPT 5.1 as the best and most insightful model, and consistently select Claude as the worst model, with the other models floating in between. But I'm not 100% convinced this aligns with my own qualitative assessment. For example, qualitatively I find GPT 5.1 a little too wordy and sprawled and Gemini 3 a bit more condensed and processed. Claude is too terse in this domain. That said, there's probably a whole design space of the data flow of your LLM council. The construction of LLM ensembles seems under-explored. I pushed the vibe coded app to github.com/karpathy/llm-coun… if others would like to play. ty nano banana pro for fun header image for the repo
101
Had a walk in the woods earlier, and the thoughts started flowing in. Didn't want to be glued to my phone though—being able to just open @AprilAssistant and voice some notes to myself was perfect. Voice has seriously hit an inflection point.
2
95
Keep your eye on MCPJam, looks like the founders are maniacal about making a great product. I reported a bug and they turned it around in record time.
3
5
121
Nice UI too. Super clean
1
63
Internet not working. Falling back to parametric memory.
59
When you're just cranking solo on a repo, it's kind of nice having @coderabbitai as a sparring partner. I really love this product! Copy/paste the vibe coding prompts is such a halo feature.
54
I just listened to Hacker News via @gethuxe this morning while making coffee. Great experience! Would it be possible to add @_akhaliq's HF Daily Papers feed as well? huggingface.co/papers/date/2…
1
2
4
663