Katsuya

So proud of this partnership 🇯🇵 I personally onboarded CMIC, the first & largest CRO in Japan, last year, and watching their unmatched domain expertise & craft meet Bluenote's AI has given us a real glimpse of what's possible for this industry. Excited to go deeper together!

Pre-moltbook: "Is this a real human?" Post-moltbook: "Is this a real AI?" We need CAPTCHA/Worldcoin for AI.

@0xMetaHulk Not sure if there's huge diff in the code quality but, for me, it's the UX: - no setup needed to run 10s of tasks in parallel (cloud vs local) - leave comments like code review and agents async address them

@kn How is it compared with cursor when it comes to code quality?

Asked Codex to internationalize our app and localize it into Japanese before bed last night. Woke up to complete Japanese support this morning 🇯🇵 What would have taken a few days was done overnight.

@SteveMoraco I broke down the task by url paths and folders and run 10-20 tasks in parallel. I had to resolve minor merge conflicts. Something like map reduce could be useful workflow here.

@kn did your prompt have to specify for it to do the whole task without asking you again? have yet to get long tasks to complete successfully

@aksmckenzie A potential workflow could be to have a GitHub Action run this on every PR so that developers don't need to think about i18n/l10n.

@kn i love this so much. are you going to run it again every time you ship a new feature?

@tanin 🤦🏻‍♂️

@kn I chose every carrier and didn't get an otp message... 😂

This healthcare portal asks for my mobile carrier, and Google Fi isn't in the list But why do they need the carrier???

@ai_sentience x.com/kn/status/1923…

@kn how long did the process actually run for in total?

@greggoriesmom @gdb It actually did it in 10-20 mins. Unfortunately I’m not AI so I had to sleep.

@gdb did it really need to work overnight on this seems excessive

Codex for internationalization:

@tuhinone @tuhinone @amiruci LFG! 🚀

It's been a wild couple of years. Thanks to all our customers who partnered with us as they bring their AI products to market. It feels like we're at the precipice of something bigger. We've raised some more capital. This isn't the important part of the company or the journey. It's support and conviction to go bigger and faster. We have a ton of upcoming launches that you're not going to want to miss.

More on our 15+ AI agent workflows, and building for the life sciences industry: bluenotehealth.com/news/lux-capit…

Domain-specific agents are driving massive productivity gains. Success requires industry-tailored UX design, evaluation frameworks, guardrails, and post-training.

Thank you to @eladgil and @avichal for their incredible support from day zero. Their advice and help have been instrumental in getting Bluenote off the ground. Excited to partner with @deenashakir at Lux Capital and the incredible investors participating in this round.

We're seeing huge demand and expanding our eng team. Bluenote automates complex workflows with AI agents and accelerates bringing medical breakthroughs to patients - far beyond the incremental gains of chatbots. Come build with us! (DM me)

GenAI use cases for cancer screening and care from @Color x @OpenAI 🚀

@nedteneva @AndrewYNg Curious to hear what your findings are though!

@nedteneva @AndrewYNg We focused on blinded subjective scoring by professional translators and our customers, and haven't benchmarked against standard metrics.

I think AI agentic machine translation has huge potential for improving over traditional neural machine translation, and am releasing as open-source a demonstration I'd been playing with as a fun weekend project. Using an agentic workflow, this demonstration (i) Prompts an LLM to translate from one language to another, (ii) Reflects on the translation to come up with constructive suggestions, (iii) Uses the suggestions to refine the translation. In our limited testing, this is sometimes competitive with, and sometimes worse than, leading commercial providers. But it gives a highly steerable translation system where by simply changing the prompt, you can specify the tone (formal/informal), regional variation (do you want Spanish as spoken in Spain or as spoken in Latin America?), and ensure consistent translation of terms (by providing a glossary). This is not mature software. But I hope the open-source community can make agentic translation work much better. Given how a simple reflection workflow already gives decent results, I think there's significant headroom to make agentic translation much better. Releasing an early software prototype like this is something new I decided to try to see if it is helpful to the developer community. I'd love any feedback on this. Thanks to Joaquin Dominguez, @nedteneva, @JohnSanterre for help with this. github.com/andrewyng/tran…

@Flynnjamm Welcome to SF!

Officially joined team SF this week ✅