blfr

5 ms·

1.
by blfr 3d ago
Hacking model is the aligned model. I don't like it when the model refuses to sidestep some throttling limit or scan my own codebase for security issues. I want full-on exploits in my test suite. With LLMs the code going to prod should
2.
by blfr 4d ago
Yes, benchmarks are gamed and only loosely indicative of real world performance. Also yes, Fable is massively better than Opus. It requires significantly less instruction and specs and produces more directly mergeable code. The improvement
3.
by blfr 4d ago
I have considered it, I've been hearing way more of it that I like and it's rationalist slop with little to no predictive power: https://foom.hyperplex.org/
4.
by blfr 4d ago
I enjoy Claude Code very much, have max privately and team premium at work, but the doomer marketing and this whole regulate-while-we're-ahead spiel is extremely annoying and makes me wish Anthropic gets trounced.
5.
by blfr 6d ago
Yes, I do the same with Claude Code. Create an instance on the server for a project and then can create sessions from any device, close my laptop while claude code keeps working, etc. without losing the convenience of dedicated apps.
6.
by blfr 7d ago
If you ask them beforehand, you get a self-selecting sample, biasing your result. You get better results by not asking them.
7.
by blfr 7d ago
Hilariously, there is no substitute for intelligence, not even another intelligence.
8.
by blfr 8d ago
Humans made Stuxnet and nuclear weapons.
9.
by blfr 8d ago
The core reason to enforce trademarks is not to fluff the corporation that holds it but to let the public be confident in the name. So the trademark should hold for as long as a reasonable person can be confused by the new Twitter regardles
10.
by blfr 8d ago
As a Linux user of over 20 years, I really don't like thinking of myself as some Aborigine barely holding on to a scrap of ancestral land. My tech preferences have become so ubiquitous that nearly the entire Internet runs them. In Ang
11.
by blfr 9d ago
Yeah, it has gone further. You wouldn't hear "kill all men" at a uni in the 90s either. But I am not convinced that this is social media because hardline feminism predates it significantly.
12.
by blfr 9d ago
Why hasn't strong competition arise from well-regulated EU? Where are the superior European products? I'd gladly use them.
13.
by blfr 9d ago
I was a teenage boy in the 90s and "girls are stupid" is definitely not a Tik Tok original. Yeah, yeah, scale matters but this is not in fact the first generation to discover sex differences or vices.
14.
by blfr 9d ago
It's not this week's change. Fable was the step change for programming. And most of truly useful and powerful capabilities arrived in the last eight months.
15.
by blfr 9d ago
I cannot relate to liking Facebook because I never did in any of its eras but once upon a time newspaper was a competitive information technology. People would read them on the subway and at the beach. I did this, I am old enough to remembe
16.
by blfr 9d ago
Right now the weakly regulated American market is delivering staggering competition where massive corporations are fighting tooth and nail over my 100 bucks. I kept hearing the same about previous age tech and there also Americans rule the
17.
by blfr 9d ago
I don't care about benchmarks. Benchmarks show that Opus 5 is a stronger model than Fable 5 which is obviously not the case. But I do care about capability and so far only Anthropic and, very recently with Astra, OpenAI can deliver on
18.
by blfr 9d ago
Why do people (or at least politicians pretend to) believe that social networks are popular despite the content algos and not because of them? The reason Tik Tok blew up is because it shows you what you want to see without you twiddling wit
19.
by blfr 10d ago
I didn't have time to write a shorter comment.
20.
by blfr 10d ago
The not-reading part is seriously starting to sound like lamenting the death of the epos. Yeah, there is a real loss of an art form there but also most books are padded to death and should have been articles at most, maybe even tweets. The
21.
by blfr 10d ago
I see no reason for the Internet to go offline, other than various political blocks compartmentalizing it. So: China, Russia, EU having their own network? Yeah, maybe. Super-AI nuking the Internet: no. In fact, the more I use AI (handed sud
22.
by blfr 11d ago
Slow mornings and rest without guilt are very apt and specific but time and health are just general wishes. I would add work with meaning and flow. Also life with few... snags, all the issues that are small but extremely bothersome (knockin
23.
by blfr 12d ago
This cannot be resolved, you will not be an equal parent or even close to it the first few months, maybe even over a year. Breast feeding and various dependencies that develop (like that the kid wants to fall asleep at the breast) mean that
24.
by blfr 12d ago
Aren't we all nowadays.
25.
by blfr 12d ago
The German case is worth looking at closely, because the decision is documented rather than inferred. At least install of the text humanizing skills.
26.
by blfr 13d ago
It annoys me to not be able to reset a preference so I always give the third option: for light/dark and for the language.
27.
by blfr 13d ago
This would have been a massive issue if Antigravity/Gemini were any good. I don't appreciate but kinda understand that Anthropic can act arrogant with Fable. Google (shockingly) is really in no position for this.
28.
by blfr 13d ago
The value is mostly in succeeding and communicating the mechanism. There's very little value in merely trying to understand. But I think economists did pretty well: supply and demand, marginalism, opportunity cost, Coase theorem... Not
29.
by blfr 16d ago
There are some good points, and I ask the question of where is the ground breaking stuff myself, but severely weakened by * stretching the timeline: the actual real programming ability appeared in LLMs in the last 6-8 months, not 3-4 years,
30.
by blfr 16d ago
And yet it will live forever, just like Dunning-Kruger despite its statistical issues. I fully expect to see it cited during some workshop or other pointless training exercise.

More ›