Hacker Newsnew | past | comments | ask | show | jobs | submit | nikcub's commentslogin

Enabling /extra-usage in my (personal) claude code[0] with this env:

    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-4-6[1m]"
has enabled the 1M context window.

Fixed a UI issue I had yesterday in a web app very effectively using claude in chrome. Definitely not the fastest model - but the breathing space of 1M context is great for browser use.

[0] Anthropic have given away a bunch of API credits to cc subscribers - you can claim them in your settings dashboard to use for this.


I also remember when the playstation 2 required an export control license because it's 1GFLOP of compute was considered dangerous

that was also brilliant marketing


Regardless of what you think of OpenClaw, Peter is a great hire - he's been at the forefront of brute-forcing app development with coding agents.

the way they respond to security and privacy incidents + publishing technical security + privacy papers / docs

Good one, yes, that is important.

Apple = Run more commercials with black backgrounds and white text that says

SECURITY

PRIVACY

---

Heyyy it never said "good privacy" perceive as you want...

Don't publicly acknowledge that you were the reason someone got murdered and 1000 VIPs got hacked.

One day when I'm deemed a 'Baddie', I looked at Apple as inspiration.


And do they approach Security as a Feature or as a Process. The fingers on one hand are enough to count them...

and with neuralink it would generate slides of the audience naked

Funny I consider this valuation modest considering what the max extent of the investment thesis is here.

SaaS and legal market caps have already contracted a multiple of the combined OpenAI + Anthropic valuations just based on the _threat_ of what they may be able to accomplish.

They'll have the data + knowledge edge over open alternatives and be able to implement + deploy (see the story about Anthropic employees being at GS for 6 months already[0])

[0] https://www.cnbc.com/2026/02/06/anthropic-goldman-sachs-ai-m...


What’s your max extent? I was just doing some napkin math to think about where they’ll cap out.

claude code is big enough now that it really needs a preview / beta release channel where features like this can be tested against a smaller audience before being pushed out.

as a regular and long-term user, it's frequently jarring being pushed new changes / bugs in what has become a critical tool.

surprised their enterprise clients haven't raised this


> In what world is public opinion not universally against the cartels? It's hard to take you seriously after that.

They definitely care about not ratting the cage with the US - they don't harm US federal agents, or take US hostages, and the last incident of Americans being killed in Mexico by cartel-affiliated gunmen in a case of mistaken identity - it was the cartel who handed the perps over and apologised[0]

[0] https://edition.cnn.com/2023/03/09/us/mexico-matamoros-ameri...


I use opus for coding and codex for reviews. I trigger the reviews in each work task with a review skill that calls out to codex[0]

I don't need anything more complicated than that and it works fine - also run greptile[1] on PR's

[0] https://github.com/nc9/skills/tree/main/review

[1] https://www.greptile.com/



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: