HN user

WinstonSmith84

1,148 karma
Posts0
Comments399
View on HN
No posts found.

I'm rather scared of US models - if Anthropic was the only AI provider in the world, it's easy to see that common people would have no access at all. Thankfully there is OpenAI which compete 1:1 with Anthropic (at a slightly lower cost) but most importantly the Chinese models keep Anthropic, but also OpenAI in checks.

And I'm saying this as someone working for American companies.

CursorBench 3.1 21 days ago

that benchmark seems to match my experience. GPT 5.5 is significantly better than Opus 4.8, last time I tried composer 2.5 it was truly dumb, and Fable to me looks to be on par with GPT 5.5 but .. different overall ... The best is to have a LLM-peer-review between GPT and Opus (now Fable) for best outcome.

Looks a little bit sensational and .. very much misleading

1- The largest donor, a16z is certainly invested in Crypto, but that's a venture capital firm, not crypto firm.

2- Fairshake PAC is basically a Crypto PAC, so it's obvious that Crypto firms are funding that PAC, right. I quote Wikipedia:

Fairshake is a Super PAC funded by the cryptocurrency industry that supported pro-cryptocurrency candidates in the 2024 United States elections.[2][3][4][5] Major contributors include Coinbase, Ripple, and Andreessen Horowitz.[6] Fairshake spent nearly twice as much on Republican candidates than on Democratic candidates.[7]

That PAC doesn't look like they are ideologically oriented, they just care about whether a candidate is pro crypto or not, whether it's dems or reps.

3- Leading the Future is not even a Crypto PAC at all, it's an AI PAC

Yes, this is actually "funny" that Anthropic feels the need to build such intrusive features into Claude Code, when anybody can build a (basic) Claude Code alternative. And the Chinese labs are certainly not "anybody". One may wonder what Anthropic really tries to achieve aside from awful publicity.

Maybe not when answers will be almost instant, but the biggest gap to me now is latency, not capability. Having to wait for 5 minutes and more for an answer, I can switch to another project / feature / etc. and run work streams in parallel.

There is no money made from these people though .. people who are using ChatGPT to plan for their next week-end or their next vacation aren't paying a $100 or $200 monthly subscription. As for non coder office workers (accountants, PMs, etc.), they use Microsoft or Google products which all integrate AI to some extent within their products - with RAG for Sharepoint to some basic AIs to generate text or automate work in spreadsheets .. the models used there are already capable enough for all what's needed (I think Microsoft is using GPT 5.1 or 5.2 in its latest iteration but for sure no GPT 5.4/5.5). The thing is, Software development is where money is made for these labs

That's hardly a speculation that he cares very much about the stock market, more than any of his predecessors. It's also why he takes a L instead of going into an adventure in Tehran. Last but not least, "it's the economy, ..." is on everyone's mind, including his

That makes this quite a bit different situation than publishing the repo on a personal account.

If that was the issue, it's not very hard to delete it either... Either nobody cares, or that repo is not as bad as for Google as it seems.

He obviously goes over the line, he tried to put GPS trackers on the cars of ministers. He "stalks" their families, and dox their children online.

I agree with you that he goes over the line, but only if these ministers are totally unrelated to the measures they are trying to impose on the population. If not, he just gives them a taste of their own medicine.

looks like to be for pentesting. If you look at their affiliates, it's a bit more clear https://lab401.com/collections/pentesting/products/lab401-pe...

But I came here too, to understand what it is about and we are probably just not the audience :-)

Now, I'd be curious to know how good these devices are and/or how insecure these RFID chips are - because 500 USD is quite cheap for a device supposed to impersonate you or open your car or your hotel room.

Claude Fable 5 1 month ago

With GPT 5.5 on the $100 plan, it's hard to hit any 5h/7d limits - while allegedly being better than DeepSeek 4 pro. Not sure why, or how you spend "a few hundred dollars of spend".

With that said, I still had the Pro plan on Claude, I didn't expect much, but it blew up my 5h allowance on Fable with one simple single prompt, and it didn't even complete lmao

If you bet against the hype and it goes on for a few more months/years you lose as much or more than if you went along for the ride.

Burry is well aware of this ...

Well, no he isn't well aware of this, apparently. He's been right in 2008 but he has been spectacularly wrong for the last 5 to 10 years, like shorting Tesla or Nvidia at the worst possible moments - and eventually closing his hedge fund...

On the plus side, their database has still not been hacked .. like it has happened one month ago in France, with the "ANTS" (the French equivalent). More than 10 millions people had their data leaked including pretty much everything from SSN, to email, phone, etc. A mine gold for Phishers and Scammers.

I can see that Visa and Mastercard are freaking out, not because Pix can take over their business model, but because it can give ideas to other countries doing the same.

I've spent three months earlier this year in Brazil and never used Pix once. Not because I didn't want, but because I couldn't, or let's be honest: my time was not worth the hassle. To be able to pay with Pix, one needs to get a CPF (Brazilian Tax ID). Then to open a bank account, mostly local banks only accept Pix, with which you can tie your CPF. It's possible but it's definitely not straightforward the slightest. All the while Visa and Mastercard work everywhere in the country, I almost never had to pay in cash, even some sellers in the streets accepted regular credit cards.

Pix is certainly great, but locally only, and if every country comes with its own system and Visa or Mastercard disappear, we are going to go back to how people used to travel 50 years ago: with a lot of USD bank notes hidden in your hotel room or elsewhere ...

Pix is a good local idea, but the world needs something better.

fun fact for Paraguay: the Itaipu Dam is one of the largest in the world located between Brazil and Paraguay, where each country gets 50% of the production. But 50% of that production for Paraguay, a country of 7 millions inhabitants, means that it cannot consume that much, so it's essentially reselling that energy to Brazil, a country with 30x more inhabitants. Paraguay only uses about 1/3 of its share (and thus resells 2/3 to Brazil).

It's just another military adventure ending in a disaster - probably the most humiliating in a long long time. But to your point, it's better for the US to admit defeat now, than in 2 or 3 weeks, let alone in 2 or 3 years. If a parallel can be made, Russia would have been best advised to have done the same 3 years ago.

Not discussing Mythos here, but Opus. Opus to me has been significantly better at SWE than GPT or Gemini - that gets me confused why Opus is ranking clearly lower than GPT, and even lower than Gemini.

I'm not following the relationship - because you'd have to pay, thus it's not "free" speech? It's hard to argue that having to pay a minimal fee (of let's say $1 per month) would be something against free speech. But the payment shall remain anonymous obviously.

An Update on Heroku 6 months ago

I used Heroku extensively before AWS reached its current level of maturity. Heroku made it incredibly easy to create cool apps. When Salesforce acquired it, and knowing a lot about Salesforce, I expected tight integration to address use cases where Apex is too limited (Apex being Salesforce's native language). There were (and still are) numerous such use cases. Unfortunately, this never materialized, and Salesforce gradually shifted away from a dev-first platform toward click-based config and heavy reliance on middleware for all kind of integrations.

It's been a butchered acquisition and missed opportunities along the way. And now it ends up just like Microsoft's Skype.

Termux 6 months ago

Right, I enabled it, and got that exact error when starting the Terminal app on my Xiaomi 15: "Non-protected VMs are not supported on this device."