If the side by side wasn’t so vertically aligned Im not sure I could tell the difference.

I’d love to use this as a base for a math model. Let’s see how far it can get through the last 100 years of solved problems

Maybe they’re just using 4o transcribe but I do wonder if they've fine tuned for medical terminology.

Comments URL: https://news.ycombinator.com/item?id=43763382

Points: 5

# Comments: 2

Do you mean the strategy or are you planning some freemium model?

I guess six years is the longest time horizon in the question but that "if" around AGI and its impact does a lot of work. This maybe assumes it'll happen sooner? Or, is it similar to the AI art prediction where the 6 year horizon is just a long enough period of time such that we'll see if we're headed towards AGI?

2. It's a mix of ego and motivation. When the work is going out to the public, for me it's just the fear people will end up viewing my work the same as they'd view a Neil Breen film.

To be fair to word2vec (rather, word embeddings) I think both require a fair amount of sentence context.

On a semi-related note, one of the reasons I avoided tackling smells yet is because so much written about smell is in the form of perfume/cologne marketing speak. Asking gpt-4o for smells lists that "the smell of jasmine and tuberose [...] evokes the mystery and elegance of a moonlit garden". I'd hope modern models would understand that this is nonsense but I can imagine a word2vec model would end up with bizarre associations.

When you refer to averaging embeddings together, do you mean averaging a bunch of sentences/words for "male" to get a general concept vector or do you mean averaging two different words, like "royal" and "adult male", to get to the combined concept, say "king"?

Comments URL: https://news.ycombinator.com/item?id=42649749

Points: 2

# Comments: 4

Comments URL: https://news.ycombinator.com/item?id=41432164

Points: 1

# Comments: 0

Maybe I'm the equivalent of that idiot fighting against JS frameworks back when they first came out it but it feels pretty simple to just use individual clients and have pydantic load/validate the output.

Dear Ted,

I hope this message finds you well. My name is Reuben Morais, and I am a Co-Founder and CTO at Coqui. Today, I have an important update to share regarding the future of our services.

As a valued member of our community, your support and engagement with our text-to-speech technology have been integral to our journey. Our mission has always been to innovate and provide the best possible service to our users. In line with this, we've made significant strides in open-source technology, contributing to the broader community's growth and development.

Upcoming Service Cessation: We have made the difficult decision to discontinue our paid SaaS web application and REST API services. This means that our servers will be going offline in 14 days, on December 11th, 2023. Active subscriptions will be automatically canceled and reimbursed for any remaining time after December 11th.

Action Required – Download Your Assets: We understand that this transition may cause inconvenience, and we are committed to making this process as smooth as possible for you. You have until December 11th, 2023 to download any assets or data you have stored in Coqui Studio. Please ensure you retrieve all necessary assets or data before this deadline, as access will not be possible afterward.

Future of Coqui TTS: While our Coqui Studio application is being discontinued, our latest XTTS model has received fantastic feedback from users in 16 languages. The model is available for non-commercial use in HuggingFace, and commercial licensing inquiries can be made by sending us an email.

Support and Assistance: For assistance with downloading your assets or any questions you may have, please do not hesitate to contact our support team. We are here to help you through this transition.

We deeply appreciate the trust and support you have placed in Coqui. Thank you for being a part of our journey.

Best regards,

Reuben Morais CTO @ Coqui.ai

Comments URL: https://news.ycombinator.com/item?id=38437743

Points: 2

# Comments: 3

That said especially in python it’s not that hard to reimplement things yourself in a cleaner way. Output parsing for agents was nice but with the function update from OpenAI it’s not really necessary (if you’re just using their API)

* Nuclear still goes nowhere, wind and solar continue to rise in use. Maybe some Elon-esque person starts experimenting with gravity batteries.

* Self driving cars still aren't a thing.

* The price war with uber and lyft ends, and both are still around just more pricey. Food delivery for regular restaurants will stop but chains and delivery only places in cities stick around.

* Commercial FM Radio will be close to death and shut down in a lot of areas.

* Silicon Valley continues it's pattern of cycling monopolies, or ya know, just buying the next one.

Cynical Takes:

* The general public in the US becomes increasingly sick/disinterested/numb as social media and blogs become ambiguously trustworthy (bots, propaganda, etc.) and and local media is swallowed/assimilated into large media groups.

* Moderate cuts to US and UK social benefits continue gradually over the decade.

Hot Takes:

* Twitter gets nationalized due to it's frequent use by politicians.

* Small scale functional fusion reactor is built/being built by the end of the decade, presumably in Europe.