TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Anthropic raising funding valuing it at $60B

115 点作者 jmsflknr4 个月前

16 条评论

onlyrealcuzzo4 个月前
Non-paywall link: <a href="https:&#x2F;&#x2F;archive.is&#x2F;ucguC" rel="nofollow">https:&#x2F;&#x2F;archive.is&#x2F;ucguC</a>
highfrequency4 个月前
Private market valuations are funky and hard to reason about. In the public markets, valuation represents the current equilibrium of supply and demand - very roughly an average of a large number of opinions.<p>In private markets, especially with the recent trend of selling a tiny portion of the company at a massive price, the valuation represents something much closer to the <i>maximum</i> that any investor in the world thinks the company is valued at.
评论 #42628856 未加载
评论 #42626870 未加载
评论 #42628938 未加载
评论 #42626389 未加载
评论 #42626478 未加载
minimaxir4 个月前
I&#x27;m far, far more bullish on Anthropic than OpenAI at this point. It has relatively less hype and media attention but better products downstream, which in the end is what&#x27;s going to be the winner in this LLM ecosystem.
评论 #42626029 未加载
评论 #42626623 未加载
评论 #42633155 未加载
erulabs4 个月前
I really love Anthropic, I&#x27;m building a business around it, and chat with Claude almost every time I go out for a beer after a long day... But having a valuation equal to _Stripe_ seems a bit insane. Either they&#x27;ve got some secret sauce or AI hype is a bit much or Stripe is undervalued.<p>Especially considering I pay for Claude, the beer, and my business _thru Stripe_!
评论 #42626482 未加载
评论 #42626508 未加载
评论 #42626513 未加载
bnchrch4 个月前
Theres two futures I see.<p>1. We create near AGI<p>2. We create ASI (Artificial Super Intelligence)<p>In the first scenario, an investment in any AI model company that does not own its own compute is like buying a tar pit instead of an oil well. In other words, this future has AI like a commodity and the entity who wins is the one who can produce it at the lowest cost.<p>In the second scenario, an investment in any AI model company is like buying a lottery ticket. If that company is the first to create ASI then you&#x27;ve won the game as we know it.<p>I think the minor possibility of the second scenario makes this a good investment. But it definitely feels like all or nothing instead of sustainable business.
评论 #42626150 未加载
评论 #42626109 未加载
评论 #42626250 未加载
评论 #42626295 未加载
评论 #42626133 未加载
评论 #42626406 未加载
评论 #42626089 未加载
评论 #42626288 未加载
评论 #42628256 未加载
评论 #42626396 未加载
评论 #42626173 未加载
评论 #42626898 未加载
评论 #42637323 未加载
评论 #42629216 未加载
评论 #42626318 未加载
airstrike4 个月前
OT: FWIW I think &quot;Anthropic in Advanced Talks to Raise $2B, Valuing It at $60B&quot; makes for a better headline and is basically a shorter version of the first line of the article. It&#x27;s a pretty factual edit so I don&#x27;t think it qualifies as editorializing.
datadrivenangel4 个月前
&quot;The startup’s annualized revenue—an extrapolation of the next 12 months’ revenue based on recent sales—recently hit about $875 million, one of the knowledgeable people said. Most of that has come from sales to businesses.&quot;<p>Anthropic&#x27;s team plans are actually pretty pleasant to use. It does seem strange&#x2F;funny though, that a significant part of the evaluation choice from a business perspective is that Anthropic is more trustworthy than OpenAI.
awongh4 个月前
Crazy that Anthropic has a fraction of the AI model market that OpenAI has, but a valuation that&#x27;s larger than their proportional share of the market to OpenAI....<p>The last valuation was at $157 billion- Anthropic is valued at 1&#x2F;3 of OpenAI but has 1&#x2F;10th of the market share....
评论 #42626268 未加载
评论 #42626102 未加载
评论 #42626369 未加载
评论 #42626528 未加载
optimalsolver4 个月前
Considering how fast open source models are closing the performance gap, this is rather optimistic.<p>Maybe if they somehow fixed hallucinations and kept the secret sauce to themselves, I could see them being worth that much, but all the top labs seem to have given up on that problem.
评论 #42626190 未加载
CobrastanJorji4 个月前
This is a great example of creating value by making sure everybody&#x27;s got a piece of the pie.<p>Say you&#x27;re some OpenAI employees. You know a lot about AI, you&#x27;ve got the resume, you&#x27;ve got some buzz, and you want your own successful startup. How do you make sure that it gets maximum preferential treatment and first dibs on all the data and GPU? By making sure the big dogs are all going to get super rich off of your success.<p>So they got billions in investment from Google, even more billions in investments from Amazon, half a billion from FTX (whoops), then some VCs for additional shmoozing power, and you&#x27;re good to go. It helps to have the technical chops and a good product to distinguish yourself, but at that point, Amazon and Google are both going to go out of their way to shove your AI into everything, so having something to contribute is practically just a nice bonus if you can manage it.
jedberg4 个月前
A lot of folks complaining about the valuation. It&#x27;s important to remember that investing in an AI model company isn&#x27;t the same as investing in most other businesses.<p>You&#x27;re not investing hoping that they turn into a big business with a nice return. You&#x27;re investing because you assume the value will either be zero or infinity.<p>If they achieve AGI first, then the valuation you invested in doesn&#x27;t matter because the value will basically be infinite (or will completely change society in a way that money won&#x27;t matter anymore).<p>If someone else achieves AGI first, the value is basically zero.<p>And if AGI isn&#x27;t achieved, well, there probably won&#x27;t be any exit. But if there is, it&#x27;s a nice bonus if it still has any value.
评论 #42637408 未加载
nojvek4 个月前
Lots of money at the top 1% looking for high risk high returns.<p>The top 1% in US own 40% of wealth, and top 10% own 80% of it. We are beyond Pareto ratio of 20:80.
deadbabe4 个月前
The best AI model will not win the AI wars, it will be the most cost effective model at scale that wins.
nsoolo4 个月前
In the last few months the quality of all Claude models has gone down a lot, I hope that in the future they plan to improve it.
评论 #42637603 未加载
评论 #42631913 未加载
graycat4 个月前
Come on, guys, help me out here: Take some people, have them write essays, collect and process all the essays, which to me sounds like one step above high school plagiarism, and now expect to discover some &quot;new, correct, significant&quot; <i>content</i> none of those people knew?<p>I&#x27;ve created some new ideas: Yes, maybe all I did was &quot;read what other&#x27;s had done and took the next step&quot;, but that step seemed <i>novel</i> to me and, thus, just NOT in the &quot;essays&quot; or &quot;what other&#x27;s had done&quot;. Uh, just how the <i>novelty</i> happened does not seem to be in any of the essays or &quot;what other&#x27;s had done&quot;?<p>Okay, maybe a two-step approach: (1) Make wild guesses. (2) Run experiments and test the guesses. But is current AI doing either of (1) and (2)? Right, for some board games can do both (1) and (2). Test the guesses with the content of the essays or &quot;others have done&quot;?<p>Here&#x27;s a simple example: At one point the FedEx BOD wanted some revenue projections, uh, <i>seriously</i> wanted as in else &quot;pull the funding&quot;. People had hopes, wishes, but nothing that sounded objective. Soooo, I noticed, guessed (1) growth would be mostly from the happy existing customers (2) <i>influencing</i> customers <i>to be</i>. The influencing would be a customer to be receiving via FedEx a package from a happy customer. So what? Okay, for time <i>t</i> let <i>y(t)</i> be the revenue at time <i>t</i>. Let <i>b</i> be the total size of the market, i.e., the revenue when do have all the <i>target</i> customers. Then at time <i>t</i>, the growth rate would be proportional to both (a) the number of current current customers and, thus, proportional also to <i>y(t)</i> and (b) proportional to the number of customers to be, and, thus, to <i>(b - y(t))</i>. So for some constant of proportionality <i>k</i>, we have that the growth rate<p><i>d&#x2F;dt y(t) = y&#x27;(t) = k y(t) (b - y(t))</i><p>which has a simple closed form solution. Then for any <i>k &gt; 0</i>, can do some arithmetic and find <i>y(t)</i> for any <i>t &gt; 0</i> and draw a graph. Do this and pick a <i>k</i> that yields a plausible, reasonable graph, and present that to the BOD. It worked, i.e., pleased the BOD which did not &quot;pull the funding&quot;.<p>This work was 100% mine. Lot&#x27;s of people in the office worked on the problem, but none of then had any ideas as good as mine -- i.e., could have them all write the &quot;essays&quot;, process those, and still not come up with the little differential equation, its <i>closed form</i> solution, or a reasonable <i>k</i>.<p>It seems to me that the essays, what others have done as <i>training data</i> just does not have or have a path to work that is new, correct, significant. Uh, can we train the AI in how to guess and test (beyond board games), how to start with a BOD request and, a description of the <i>why and how</i> of business growth, some calculus and get an answer, take epicycles and come up with <i>F = ma</i>, Tesla&#x27;s experiments, Stokes formula, and get Maxwell&#x27;s equations, make a wild guess and propose the Michelson-Morley experiment, get <i>E = mc^2</i>, use the inverse and implicit function theorems, Riemann&#x27;s work on manifolds, and get general relativity, solve Fermat&#x27;s last theorem, make real progress on the Riemann hypothesis??? Uh, in short, we need a <i>idea</i> not in the <i>training data</i>? Soooo, need to train the AI to have <i>ideas</i>? <i>Ideas</i> are just using what&#x27;s in the essays to make connections in a graph and then exploring the graph until get a path to an answer?? How do such training? Can process existing text yield such training data?
eGQjxkKF6fif4 个月前
Meanwhile I use the AI on one of my projects and it cuts me off making me wait 3 hours. I get it, but for $60Billion I think they can afford some IO.
评论 #42630530 未加载