5 ms·
The funding frenzy at Anthropic
- bookofjoe 3y agohttps://www.nytimes.com/2024/02/20/technology/anthropic-funding-ai.html?unlocked_article_code=1.XE0.hDv8.cbHf2F3GXEh1&smid=url-share https://www.nytimes.com/2024/02/20/technology/anthropic-fund... https://archive.ph/SGpyY https://archive.ph/SGpyY
- highfrequency 3y ago> Some investors have questioned such deals because companies like Google and Amazon are investing money that ends up bolstering their own revenues. This is brilliant. It's a win-win for cloud providers like Microsoft and Amazon to invest billions into AI companies because the "investment" is guaranteed to flow back in as cloud revenue (on which they have huge valuation multiples) regardless of the AI company's fate or eventual profitability. The AI companies then get to quote massive valuations based on these investments, even though the investments were made under very unique circumstances (ie "most of this investment will flow back to us as revenue"). Finally, other investors[1] who do not benefit from the cloud spending and who may be less discerning on price will FOMO in at whatever the current valuation is. These investments not only have no stipulations about how the funds are used, but often aren't used for the business at all - instead directly paying out founders/employees. It's a brilliant triple whammy - the technology itself is super impressive, cloud providers are naturally motivated to pump up baseline valuations due to the symbiotic relationship of exclusive cloud spend, and these valuations are taken at face value by non-cloud-provider investors eager to get in on AI hype at whatever the prevailing price is. [1] https://www.theinformation.com/articles/thrive-capital-to-lead-purchase-of-openai-employee-shares-at-80-billion-plus-valuation https://www.theinformation.com/articles/thrive-capital-to-le...
- Mistletoe 3y agoYes we’ve traditionally called it a Ponzi scheme or pyramid. We are in the part of the cycle where no one actually has to make money from AI. We’ve been here before with Uber and WeWork. All these free queries aren’t free and run on incredibly expensive hardware. The final endgame is a Dyson sphere around the sun so that we can ask ChatGPT-XX if it’s raining outside and the owners of the AI and chip companies blast off to their colony on Mars.
- lotsofpulp 3y agoHow is it a Ponzi scheme or pyramid?
- datadrivenangel 3y agoYou use later stage investments to pay off the earlier investors. Eventually, you go and IPO and the general public cashes out the later investors and then the stock goes to zero.
- lotsofpulp 3y agoIs the plan or expectation of the stock to go to zero (or otherwise dropping precipitously)? If not, I do not see the value in labeling it a Ponzi scheme. If there exists a business plan and the bet is to invest in something that provides utility, then it seems like any other investment in any other business.
- bryanlarsen 3y agoThat's what caused the 2002 crash. 99% of the revenues for internet companies came from other internet companies. When the easy money started drying up, revenues collapsed.
- reactordev 3y agoWe'll just increment the version number at the end of Web like we always have.
- blueblimp 3y agoI wonder what's happening to all that money. Back when they originally released Claude, they were second only to OpenAI as far as chatbot models were concerned. Although Claude wasn't as smart as GPT-4, it had a more pleasant writing style, and Anthropic later released 100k context. At the time, I expected Anthropic to be the next company to release a GPT-4-level model. But since then Claude has been passed by Mistral's mistral-medium and Google's Gemini Ultra. More concerningly for Anthropic, each subsequent release of Claude has actually performed _worse_ on the Chatbot Arena Leaderboard. (Claude-1 outranks Claude-2.0, which outranks Claude-2.1.) The reason for the decline in ranking is seemingly that the most noticeable update is to make the model refuse more requests. In an additional blow, the needle-in-a-haystack independent benchmark revealed that Claude's long context is not actually used effectively by the model. All-in-all, Anthropic is not looking in a good spot, despite the massive investment. They need to start releasing legitimately better models, or risk irrelevance.
- patrickhogan1 3y agoAt the MIT event, Altman was asked if training GPT-4 cost $100 million; he replied, “It's more than that.” Training costs are decreasing, but whenever there's an update to GPT-4 (e.g. training data cutoff updated to December 2023), it means the model has been retrained. The compute costs of training a model like Claude 2 are significant. Also, keep in mind that not every trained model becomes ready for production. Some are discarded, similar to how you might burn a few cookies while baking.
- rolisz 3y agoI don't think they always retrain models from scratch. Sometimes they might do continual learning (take the old model and train it on newer data)
- levocardia 3y agoSeems like the Occam's razor hypothesis would be "company that split from OpenAI over safety concerns is overly focused on safety at the expense of performance."
- HlessClaudesman 3y agoAnthropic is still not available in in EU. What's the holdup? Do they aspire to be a global service or a regional one?
- 123yawaworht456 3y agothat would be unsafe and unethical.
- vitorsr 3y agoIs it not available through Amazon Bedrock?
- vessenes 3y agoBoy I just can't stop thinking about that weekend when Sam got booted from OpenAI, and Anthropic turned down the offer to acquire OpenAI. If I were CEO of Anthropic, I would have had one job that weekend. I am very curious about Anthropic's strategy right now; they launched with "More progressive and careful about AI risks," which in my opinion is a good AI worker recruiting strategy, but not a good commercial strategy, or even a good quality strategy, based on Claude's current state. They have done some outstanding research this year -- in particular, I loved their work on tracking conceptual bundles through different layers, and I'd say their paper on embedding adversarial behavior in LLMs was my favorite paper of the year -- But, I'd like to see that research turn into more capable models. As I write this, I wonder what I'd do if I were in charge there. I think I'd probably try and one-up Mistral, and release larger more capable models more quickly. I'd try and buy Groq and build a custom vertically integrated stack that's 5x faster in tokens/s; worst case Amazon could help with the acquisition and deploy Grok into AWS. Basically the moat would be deployed tokens/s at or above certain levels of quality. I'd stop running anything but my own models on Groq hardware. There are two next-gen directions for inference in the cloud, high quality multi-modal foundation models, which are definitely coming, but also clearly need some UI work to get into a compelling product format, and very long context length models, so I guess I'd possibly be trying to leapfrog there. It's highly competitive in the model space right now, so a model-only strategy isn't good. Alternatively, I could sit back and relax and let AWS pimp Claude while we just scaled it up to 250k tokens and used MoE to get non-exponential scaling / possibly explore an SSM / Mamba architecture, and not worry too much since we have 4bn in the bank. I wouldn't be surprised if that latter path is the one they're on, sadly.
- swyx 3y agoturned down what offer? was it a serious offer? i doubt anthropic had the funds to actually acquire it.
- hackerlight 3y agoThe board approached the Anthropic CEO (Dario Amodei) and proposed a merger with him as CEO. Possibly because he presents as safety conscious. He declined.
- aduffy 3y agoThe quality of discussion in the NYT comments is exceedingly bad. In our industry it’s easy to forget that the average person doesn’t understand what’s happening currently in AI and fears the worst, based on the FUD being shouted by mouthpieces who have something to gain by being the gatekeepers of who uses AI and how.
- Animats 3y agoComing soon, the AI collapse. There's been real progress in AI, but way too much money is going in too fast for the amount of revenue generated.
- fapi1974 3y agoWe've seen this before. Lots of investments in the '99 runup had ad buys associated with them which did the same kind of double counting. The example would be Yahoo investing in some startup at an inflated valuation, paying for some or all of the investment with ad inventory priced arbitrarily and which was, for Yahoo, free to produce. Lots of things led to the 2001 crash, but valuation distorting shenanigans like that didn't help. To the extent it feeds through to Microsoft, Google, and Amazon valuations the AI dynamic could follow a similar pattern.
- tomaytotomato 3y agoMaybe I could setup a rival company called Misanthropic and secure lots of funding from AI apocalypse fantasizers?
- nojvek 3y agoAPPL and MSFT got their start in 1975 and 1976. Since then they’ve been growing and now they are the biggest ~3T jaggernauts. My biggest fear is by their sheer momentum they keep on getting bigger, capturing regulators, acquiring companies, doing cloud credit deals. They will be the technology dictators like Putin. US anti-trust is quite weak. Lina khan hasn’t had much of a win. If OpenAI cracks AGI and MSFT has the sole license to extract the most profits and set the regulations, we are in a tough spot as humanity. If Apple and Google keep their duopoly on the OS that most people use and every transaction is rent extracted at 30%, then that is a sad future. Is there room for a new player to climb the ranks and be top 2? Google and Meta are almost set for life as tougher they have a massive duopoly on online ads.