54 karma · joined January 13, 2022
Seems like an excessive announcement just to get a little extra time.
For these companies, is your argument that “pacing the frontier” is their attempt to be nationalized and protect their investments?
Red flags:
- Amodei claims RSI is here and links two sources, but neither reinforce this notion. It’s heavily agreed upon true RSI is the danger sign here (also by Coxon). Nothing indicates there is something these labs have developed that supports RSI today or even soon.
- Real US orgs are moving, like today, to Chinese open source AI run locally due to cost constraints. Ask anyone in the industry in an infra role. Frontier AI is just way too expensive to justify for coding. The big 3 labs are scared of this - their revenue and moat disappears once OSS models are progressing enough for developers to use. Kimi/GLM/DeepSeek Flash/Qwen are usable today on hosted hardware for actual coding work.
- These labs would all benefit from nationalization if their funding model fails (an easy bailout).
- Related to the above, with federal guidelines in place (i.e. restricting access to certain organizations, or establishing something like FedRAMP approval for AI use), their path to government/contractor/large U.S. company distribution becomes so much easier. They can control the supppy chain of AI for organizations.
- Preparing for (or in xAI’s case, launching) a public IPO creates excitement among consumer investors when they think this tech is as powerful as it is claimed.
- The HuggingFace hack demonstrated no new capabilities that were not previously seen before - previous models were misaligned, but weren’t as harmful as agent swarms. Anthropic’s marketing directly hopped on this bandwagon when it saw the outcome of OpenAI’s blogging making it to mainstream media (and they constantly paint China as “the” bad guy without further elaboration or research that you’d expect on APTs).
Amodei is a CEO at the end of the day. It may be 25% earnest speculative concerns, but you’d be foolish to take it at face value given the patterns we’ve seen from Anthropic, xAI and OpenAI to date. Until we see real proof of RSI we should be skeptical.
:O
The second is gonna be a lot harder to enforce, as we soon (and probably already) don't know who we're talking to on the internet - a real person or someone's agent? Will calling spaces "human only" later be seen as discriminatory by agents? How will we actually enforce "human only" spaces? Will websites like HN start to provide an "agent only" discussion forum or filter in addition to the "human only" sections?
This article, covering elements of writing like content, language & grammar, style, and citations, is verbose and is intended for Wikipedia’s army of content editors.
Walls of text are boring. Games are more fun. I turned Wikipedia’s quiz into an interactive game. I failed miserably.
This made me think this was satire/ragebait. Most important relationship?!?
RIP, another one to the Google Graveyard.
Yes, we’re making this data not valuable by design, since it cannot be linked back to the originating user to report browsing history, etc.
I ended up using the “casino” analogy to describe the new feature. Same thing, different words. The security knowledge stack is so dense that it helps to embrace analogies when explaining concepts to customers.
By “blinding” ourselves, we’re signaling to users that we’re making their data invaluable for marketing, analytics, or third party sharing. Think the EU “Reject unnecessary cookies” banner but on steroids.
Sometimes this is necessary, but often it's not. This can lead to your data being collected unnecessarily - placing your full trust in the service provider.
I studied Cloudflare’s Privacy Pass framework and implemented a modified version of it in my application Ward. As a result, Ward users can no longer effectively be tracked or identified at the app level for most requests.
Would love to hear from others’ perspective, whether you’re an security expert or everyday user, on this approach.
The average consumer is more privacy conscious than we may think. It’s our hope that architectural decisions like this early can help build and gain user trust earlier, especially in B2C products.
It is best in its class, but trips up frequently with complicated engineering tasks involving dynamic variables. Think: Browser page loading, designing for a system where it will "forget" to account for race conditions, etc.
Still, this gets me very excited for the next generation of models from Anthropic for heavy tasks.
TL;DR be cautious of the Chrome ecosystem, vet authors/publishers, and use a gut check for "free" stuff
-> 2.5 Flash Lite is super fast & cheap (~1-1.5s inference), but poor quality responses.
-> 2.5 Flash gives high quality responses, but fairly expensive & slow (5-7s inference)
I really just need an in-between for Flash and Flash Lite for cost and performance. Right now, users have to wait up to 7s for a quality response.
There has to be a better system. Maybe a public extension safety directory?
I'm bootstrapping and covering LLM costs for Ward's first couple hundred users (got about 50 users at the moment) to improve it. We have a local mode for added privacy and are dipping our feet to gauge biz. interest (client-side phishing protection is unparalleled).
I've recommended it to friends, colleagues and loved ones. I dogfood my own product, and it even surprises me every day how much more mindful it makes me of my browsing of harmful content. Would love to get feedback and testers from HN.
On the Chrome Web Store -> https://tryward.app
The real value today is in sustainability & impact - something you can’t as easily put a $ on.
We aggregated half a dozen plus disparate data sources to create a comprehensive infrastructure map of the PNW power grid. Our goal is to be able to query for and provide informed answers for grid operators, investors, and other energy adjacent businesses in the space.
(For reference): The PNW has the most abundant clean power in the US and is one of the markets with most opportunity as our consumption increases with AI.