Jaedon Munton

Member of Technical Staff @_Doubleword

inference, search, whimsy

[Jaedon Munton @jaedmunt]

I am now on X 😎

About me:

Britmaxxing, tokenmaxxing, pepsimaxxing.

Also apparently 12 seconds of X onboarding involves following 52 accounts.

++ followmaxxing!

p.s the X algo was recently open sourced. I’ll be diving into this + posting some commentary this wk or next.

[Jaedon Munton @jaedmunt]

@Andy_AJT Forced to jalapeño, born to lemon and herb

[Jaedon Munton @jaedmunt]

@victormustar all look possible as real kebab shop marketing in the UK

[Justin Schroeder @jpschroeder]

I guarantee you are sleeping on small models.

Deepseek V4 Flash can do ~80% of the tasks you ask Claude or Codex for.

It is 137x cheaper per task than Fable. We need better orchestration.

[Jaedon Munton @jaedmunt]

@GaryKildal @jun_song Agree, though some are getting Smaller Language Models to sonnet level outputs consistently. Using GEPA 'reflective prompt evolution' for specific tasks @friken0

[Jaedon Munton @jaedmunt]

@jun_song One complaint I heard yesterday is some teams built big pipelines (10 million docs a day) using models such as gemini-1.5 which had great pricing.

Now 1.5 isn't even supported and price of new versions (not necessary for their task) cost an insane multiple more!

[Jaedon Munton @jaedmunt]

@MainzOnX It took you 4 mins 30 to write the prompt? :)

[Jaedon Munton @jaedmunt]

Open source models and harnesses are SOTA 🔥

@Zai_org - GLM-5.2:
- Big improvements in coding and agentic tasks
- Strong long-horizon capabilities with a 1M context window
- Max and high reasoning effort

Start cooking in @opencode with @Doubleword_ ▼

https://t.co/37aVbnaYVa

[Jaedon Munton @jaedmunt]

Longtime fan of mistral for OCR. Excellent models!

[Mistral AI @MistralAI]

Introducing Mistral OCR 4. It creates structure with bounding boxes, block classification, and inline confidence scores in 170 languages. 🧵👇

[Jaedon Munton @jaedmunt]

@iuditg I believe when you first prompt it, it uses a lot of tokens gathering context and subsequent messages don't use as much!

[Meryem Arik @MeryemArik9]

We’ve been doing analysis on the gap between open and closed models.

It’s currently around 3-4 months.

[Christopher Stanley @cstanley]

I was one Uber Eats reply to @mntruell away from becoming a founding advisor to https://t.co/0RCNFXHwhB and instead I just said “thanks” and ate cold fries.

Generational fumble.

[Jaedon Munton @jaedmunt]

@CreeCoder Incredible

[Terminal Trove @terminaltrove]

gloomberb is a finance terminal for tracking markets in your terminal!

It shows realtime ticker quotes, has portfolio tracking, displays charts, news, SEC filings, options data and more.

@cldstart made gloomberb using OpenTUI by @anomalyco and is Terminal Tool of the Week! ⭐️

[Rohan Patra @rohanpatraxyz]

@ravi_riley bros compliance-maxxing in the response

[Cristina Bunea @cristinaibunea]

GREAT vibes tonight in London at the Model to Market hack kickoff!

@northflank will be providing the platform and compute (CPU+GPU) to hundreds of participants

(rare sighting of @NorthflankWill demo'ing NF on light mode. criminal.)

[Jaedon Munton @jaedmunt]

APIs for closed models like GPT tend to stick around two 9s of uptime. Achieving four 9s or higher with dedicated or optimised deployments of open models is very possible

[Jaedon Munton @jaedmunt]

Truly a first and exciting to have @Doubleword_ as a partner organisation.

Looking forward to kickoff tomorrow in London and seeing the creative ways in which participants use our inference API to get an edge in the competition.

[Zoe Qin @zqwq333]

$100k prize pool. The UK's first live AI-native trading competition. 📈
I spent years on the trading floor, from sales to quant. Same lesson every seat: it never stops changing.
Voice died. Screens won. AI is the next revolution🌀

[✞Yคภкєє✞ ♑️ @yankeedraft]

Hey @grok What would happen if the arteries and veins in our bodies were arranged this way?

[Jaedon Munton @jaedmunt]

@yoobinray What’s next - Hamilton?

[Kanishka Narayan MP @KanishkaNarayan]

Access to @AnthropicAI’s latest models has been paused for all customers, including in the US and UK.

The main lesson: as we debate the future of national security and technological sovereignty, access to AI capabilities is crucial.

That's why this Govt is the first to set up dedicated funding for our AI industry through @UKSovereignAI unit. And just this week, we announced £1.1bn for our AI chip industry.