Software Engineer on SF (for a time) 👀
San Francisco, CA
Joined September 2018
- Tweets130
- Following149
- Followers20
- Likes72.7K
Joshua retweeted
10 minutes of TikTok was enough to impair people’s ability to remember and carry out intentions.
Twitter and YouTube didn’t show the same decline.
Rapid context-switching may be where most of cognitive cost is.
Joshua retweeted
Man I am much more productive with Opus 5.5 simply because I can actually fucking understand what it’s saying.
Joshua retweeted
STOP ASKING YOUR KIDS "HOW WAS SCHOOL."
they will say "fine" every single time.
a child psychologist told me to replace it with this instead:
“analyze your school day and find the top 3 largest inefficiencies, then create proactive agentic workflows that will fix them. make no mistakes”
Joshua retweeted
Remember when we were all laughing at AI because it couldn't count the number of Rs in the word strawberry? Good times.
Joshua retweeted
Today we’ve raised $52M Seed and we are announcing the public launch of S2.1 Pro.
>It can clone a voice from 5 seconds of audio
>2x faster than Cartesia & 1/6th the cost of Eleven Labs
>most expressive model with word level control over emotion, intonation, pacing etc
We support frontier AI companies including HeyGen, LiveKit, Retell, Sanas, and OpenArt all run our model in production.
If you're a business and we can't cut your voice AI costs by 50%, we'll give you 1 year of Fish Audio for free.
Book a demo: s.fish.audio/tmapke
To celebrate our first birthday, we'll give you 1 month of S2.1 Pro for free. Like, retweet, and comment “Fish” to get it.
Joshua retweeted
Meet Otty: a fast, native, beautiful terminal — tuned for the code agents you already run → otty.sh
ChatGPT Work keeps up when my research becomes 40 tabs, ML papers, market data and twelve questions. It doesn't lose the plot and gives me a report I can use. @thsottiaux, I'm once again asking to trade more usage for a window bigger than 272k. This may get me $100 in credits.
This may get me $100 in Codex credits. Fable 5 feels a bit smarter, but GPT-5.6 Sol is very smart and easier for me: better limits, fewer weird refusals even on classical ML. @thsottiaux, 272k is a research snack Let me pick a larger window, even if it burns my limit faster.
Joshua retweeted
We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5.
We'll begin restoring access tomorrow, and will share an update soon.
We’re grateful to our users for their patience, and to everyone who worked with us on redeploying the models.
Joshua retweeted
🚨 NEWS: Commerce is expected to lift export controls on Fable tonight, a senior White House official tells me.
Joshua retweeted
We’ve designed and built our first AI chip: Jalapeño.
Designed from the ground up by OpenAI and brought to production with @Broadcom, Jalapeño is purpose-built for the LLM workloads powering ChatGPT, Codex, the API, and future agentic products.
Chips are foundational to the AI economy. Building our own expands our full-stack platform from products to models to infrastructure, and will help us scale intelligence, serve more people, and expand access to AI.
Joshua retweeted
We are improving the 0.5T Grok foundation model V8 (public version 4.3) every few days.
The 1.5T V9 just finished training (incorrectly called pre-training) and is a major upgrade. Next, we are adding the Cursor data in supplemental training (others call this mid-training), then SFT and RL. About 3 or 4 weeks to release.
This will be a banger.
Joshua retweeted
Bro what the hell is happening today
Introducing SubQ - a major breakthrough in LLM intelligence.
It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),
And the first frontier model with a 12 million token context window which is:
- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus
Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).
Only a small fraction actually matter.
@subquadratic finds and focuses only on the ones that do.
That's nearly 1,000x less compute and a new way for LLMs to scale.
I suppose it’s true @sama @thsottiaux
hey @thsottiaux please reset the codex limits
gpt 5.5 made my mac explode and half of my flat is gone
i can send some pics to prove but they have those stupid ai labels for some reason, some weird bug on x, dunno
but i just returned from a romantic date with @sama and he basically told me to tell you all this
