Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Amazon will invest up to $4B in Anthropic

anthropic.com

Post reply on HN
Print view View original

251–260 of 275 posts

Re: Amazon will invest up to $4B in Anthropic

#251

Post by woah »

swyx wrote: post #213
nickelbob wrote: post #184

Earlier quoted context omitted.

Please do - not trying to be snarky, just legitimately interested.

each of these are known >$100m opportunities: Generative Text for writing - Jasper AI going 0 to $75m ARR in 2 years Generative Art for non-artists - Midjourney/Stable Diffusion Copilot for knowledge workers - both GitHub’s Copilot X and “Copilot for X” Conversational AI UX - ChatGPT / Bing Chat, with a long tail of Doc QA startups (shameless plug from https://www.latent.space/p/agents )

Do they have a moat though? When MS Office integrates a more powerful model than is even available to Jasper right into the word processor, what happens?

Re: Amazon will invest up to $4B in Anthropic

#252

Post by nojvek »

riku_iki wrote: post #238
m3kw9 wrote: post #162

Looking back we can see how savvy Microsoft was to invest in OpenAI so early

future will tell, there is a chance OpenAI will just burn through the 10B on training without creating revenue stream

That sweeet seeet GPU money for Nvidia and OpenAI employees. And I guess a huge chunk of that is Azure GPUs which probably cost MS a lot less.

Cloud providers have margins >50%.

Re: Amazon will invest up to $4B in Anthropic

#253

Post by ldjkfkdsjnv »

tinyhouse wrote: post #169
doubtfuluser wrote: post #48

So what will this mean in relation to their recent Alexa announcements? are they using Anthropic in this announcement? Is it so difficult for a big company to build their own?

Alexa has been building their own LLM stuff for years now. Their models are also available on AWS. This Anthropic investment sounds much more of an AWS investment. If the Alexa LLMs were popular and capable like the best models, Amazon wouldn't invest in Anthropic. Currently if you want to use OpenAI models with better rate limits, you go to Azure. Since openai models are the most popular rn, that creates a problem f…

Does azure have better latency than directly calling openai?

Re: Amazon will invest up to $4B in Anthropic

#254

Post by jazzyjackson »

edgyquant wrote: post #112
amanzi wrote: post #41

Earlier quoted context omitted.

You didn't say which version of GPT you prefer it over. My experience of Claude hasn't been good at all - I get obvious hallucinations all the time. Whereas I get pretty consistent accurate results with GPT-4. I just tried again now and asked it about HTMX and Django since I had just been reading an article about it. Claude invented a package called `htmx.django` and a decorator that could be imported called `@hx_req…

I’ve been using Claude for months now, as has my dev team and a number of other people I know, and we don’t have this problem at all. In my experience is hallucinates less. This is compared to ChatGPT with GPT4. Anecdotal evidence is worthless it seem.

hallucination is stochastic.

Out of 1000 people flipping coins over and over again, on average people get a mix of hits and misses - there will always be someone who just get 10 heads in a row and someone else with 10 tails in a row, this corresponds to some users being wowed by the superintelligence nailing the answer every time and other users getting made up citation after made up citation.

Re: Amazon will invest up to $4B in Anthropic

#255

Post by pests »

benreesman wrote: post #145
throwbigdata wrote: post #135

Earlier quoted context omitted.

Yes, you are. Now, what are you talking about?

The broader point is that I think a lot of successful tech figures get a pretty unfair rap on HN. Few if any of these folks are stirring candidates for the Nobel Peace Prize, but Bezos or Mark or Elon or even Gates being uniquely bad? Altman, with the full weight and backing of YC in its prime and a clear first-mover advantage got bailed out of the Loopt debacle [1] in a deal with Green Dot that maybe scraped above t…

I completely missed the Socialcam era apparently. First I've heard of it.

Re: Amazon will invest up to $4B in Anthropic

#258

Post by mensetmanusman »

janalsncm wrote: post #227
nostrademons wrote: post #204

Earlier quoted context omitted.

Generative AI is very good at spewing bullshit. Unfortunately, the upper echelons of society run on bullshit. And so that's exactly why generative AI is revolutionary. The killer app is in separating dumb billionaires from their billions, just like crypto's killer app was. And just like crypto eventually filtered down to separating dumb retail investors from their thousands, generative AI may eventually filter down t…

From personal experience, I’ve saved hundreds of hours with ChatGPT already. It can piece together large neural nets in PyTorch that would take me an hour or more to get right, or annoying pieces of code which I don’t feel like correlating the disparate pieces of online info for. Of course I will review it and unit test it to make sure it works, but I honestly make more mistakes integrating it in my larger project th…

This is my experience. What does it mean for the world when skilled programmers are 10x-ing their productivity?

Is it just more vacation?

Or do things change more profoundly?

Re: Amazon will invest up to $4B in Anthropic

#259

Post by HerculePoirot »

Yizahi wrote: post #58
baq wrote: post #3

If this means alexa isn't a pet rock that can only reliably set timers, good for them.

Everyone has already moved to the much superior Alex. https://www.youtube.com/watch?v=W0_Tt0En7v4

I hope the downvote you received for this comment did not leave you ... emotionally damaged.

Re: Amazon will invest up to $4B in Anthropic

#260

Post by rnk »

mensetmanusman wrote: post #258
janalsncm wrote: post #227

Earlier quoted context omitted.

From personal experience, I’ve saved hundreds of hours with ChatGPT already. It can piece together large neural nets in PyTorch that would take me an hour or more to get right, or annoying pieces of code which I don’t feel like correlating the disparate pieces of online info for. Of course I will review it and unit test it to make sure it works, but I honestly make more mistakes integrating it in my larger project th…

This is my experience. What does it mean for the world when skilled programmers are 10x-ing their productivity? Is it just more vacation? Or do things change more profoundly?

It means that there will be higher productivity expectations based on using this. If you can't use it very well you won't be as rewarded, you probably won't be compensated as well.

We have a long human and also programming history that as we gain capabilities, the expectations of that being a standard thing grows.

Post reply on HN