I believe that comparing LLMs with traditional deterministic software is fundamentally misleading. It is extremely difficult to truly interpret what LLMs do internally, and as of now, nobody fully understands it. Even if you trained the LLM yourself, there is no source code you can simply read and learn from.
Sure, having information about how these models were trained is helpful for reproducibility, but it is basically impossible for anyone without substantial capital and access to the same (likely copyrighted) data to reproduce the model. For normal users, owning the model weights essentially means owning 100% of the model, you can inspect and study the weights in much the same way as the lab that produced the model can, you can modify the weights, and you can use and distribute them if the license allows you to
I strongly agree with antirez. I believe that an intellectually honest programmer should recognize where their knowledge and experience is mostly beneficial, and at this point (and we've seen clear signs for some time) coding is largely solved. Which doesn't mean that you can just prompt at a very high level, but that iterating with the LLM over decisions, designs and tests, will essentially allow you to not write a single line of code by hand. Software engineering is not just coding, and it involves a whole set of other tasks requiring direction and creativity that can greatly influence the quality and impact of the software.
The most common arguments against this view seem to arise either from ideological resistance, which I understand given how painful it can be to see one’s job at risk or an important part of one’s identity taken away, or from generalizing a small number of experiences with LLMs to the technology as a whole. In the latter case, those experiences may also be heavily conditioned by the user’s inexperience in working with LLMs, or by the specific use case in which they were applied. There are still certain tasks that not all models can handle reliably as of now, however some can (usually the most expensive), and they will likely continue to improve over time.
"coding is largely solved" and "self-driving is largely solved" smell exactly the same to me.
ETA: I bristle deeply at your idea that anyone who doesn't agree with you is letting their emotions get the better of them. Perhaps the world is not as black and white as you're painting it?
What does it mean for coding to be solved, and software engineering to not have been solved? If it implies there's now a set protocol that can be followed to reach good results, how has that not been the case before? If it means that we can reach better results than before, then how do we know that can't be improved further? Or does that mean we should know how English language, instead of computer code, maps to computer instructions?
Actually, this paper shows that the LLMs create a lot more slop and degradation of the code after the first prompt. https://arxiv.org/html/2603.24755v1.
So iterating with the LLM will simply make it to produce worse and worse code.
This is something we have known for a very long time, and companies are not trying to hide that either. They do it to avoid letting competitors train their models on the CoTs
Yes hasn’t this been around since Opus 4.6? I very much recall this change happening around January or February, and it was very explicitly to prevent distillation. Sonnet does not have this limitation.
Fun fact: if you go back to the old school from 2 years ago and provide explicit CoT prompts, you get the full thinking prompts back again!
So you disable thinking altogether, and instead make thinking part of the regular prompt by prompting it:
“Before providing your answer, think step by step. For example:
The use is asking me to…
I need to think about the blah blah. First, I should foo the bar, and then blah blah.
Answer: <put your final answer here>”
And tada.wav we have CoT as it worked in the GPT3 era back again.
I thought this was considered best practice? I actually prefer it to exposed thought channel, much like how I would prefer a human answer with supporting logic instead of an explanation of their problem-solving approach.
Yes, this is best practice, especially if you have a problem and can guide it a bit how to think it through. But people don’t realize that “enable thinking” literally means that Anthropic prompts Claude for something similar, tells it to wrap it inside <thinking> tokens, and that’s it.
I also don’t believe Chinese LLM labs don’t know this, so I’m fairly certain the whole summarized thinking isn’t preventing them from distillation.
lol I’ve been using this since the IRC days I think, I’ll never forget that sound; as a matter of fact, I’ve got a Claude Code completion hook that plays this sound whenever it’s done.
I had checked that page not long ago, and as far as I remember there were many "red" or "orange" days in the past 3 months. Now it's all green. That's concerning
There have been plenty of cases like this over time too. Company makes controversial change. Company rolls it back after outrage. Company slowly shifts over time until they've restored what's essentially the original controversial change.
When a company tells you their intention by announcing a change, it's often a good idea to listen. Even if their PR department does some good cleanup work in the aftermath.
Yeah exactly. When a company announces some money making scheme and it gets backlash they don't think "oops that was a mistake we won't do that"; they think "oops that was a mistake - we'll have to do it in a way that gets less backlash".
Another recent example is GitHub charging for self-hosted CI. They backtracked, but they're still going to end up doing something. They kind of have to because of all the "get 10x cheaper actions runners by changing one line" people.
I had checked as soon as I found out about the news the other day and it was there. I just checked on wayback machine and you're right, it was removed for some time.
However, if they're willing to put back that claim immediately, I doubt that their intention was to drop the free plan anytime soon, but probably it was to incentivize people to use the paid plans. Enshittification must happen sooner or later afterall, but fortunately vaultwarden exists and the export feature is highly unlikely gonna be removed immediately as the free plan disappears, so people could just switch to a third-party or self-hosted backend as soon as that happens.
I dont think its an over reaction. It's pretty common to lock in users by removing or imposing cost on exports. Having an export from today is a lot better than having nothing in 5 years when bitwarden disables exports
I've had the argument so many times with eng managers about how this password manager or that password manager will get hacked or get enshittified and I've been right 100% of the time.
Can you name a single password vault that has removed the ability to export, I would say it is a bit of wild speculation to assume this would happen. Even more so as there seems to only be anecdotal and speculative evidence this would happen.
Between the law suits, and the brand damage, there is likely very little upside for a company entertaining this idea.
It is not an overreaction at all to them replacing the principled leader who promised things with the vulture leader whose job and job history is primarily to enshittify things and sell them off.
I need to:
- Replace the extension and login on all of my browsers on all my devices
- replace the desktop application/app on all my devices
- go through and rework all the scripts that I use to automatically pull passaords from Bitwarden using the API and hope that the replacement has a good API
Nah, I think I'll stick with and keep paying/supporting Bitwarden.
You must be getting a different version of that page than me. The free tier is there but there’s no “always free” verbiage. There is “start free” verbiage.
Edit: “always free” was hidden under a collapsed section
LOL.. you are correct. Funny thing though... the 'Always Free' text is linked to a "/start-free/" action\page. One could argue that they are hedging their bets.
Some other commenter says there are Archive.org cached versions with "Start free" instead of "Always free", so they must have backpedaled on this. Maybe they realized they turned the knob a bit too much towards "hot", increasing the temperature of the proverbial water too noticeably.
I’m not willing to check all the pages on archive.org but for sure a month ago they had a big “Basic Free” tile in the plan comparison. Now it’s just Premium and Family. They are definitely downplaying the ability to use it for free.
Seems like they want to downplay the mentally that you would never benefit from an upsell to the paid plans, even if the free plan itself stays always free
Except for guessing the right continent (not that remarkable), mine is so majestically wrong that I would either dislike or fully hate all of the products I got recommended.
You can make an LLM sound very natural if you simply ask for it and provide enough text in the tone you’d like it to reproduce. Otherwise, it’s obvious that an LLM with no additional context will try to stick to the tone the company aligned it to produce
Agree. I keep asking LLMs to tell me some jokes from time to time, but never once I've found it's funny. For me, when I find myself burst out laughing from LLMs joke, I'd know we've reached AGI.
Sure, having information about how these models were trained is helpful for reproducibility, but it is basically impossible for anyone without substantial capital and access to the same (likely copyrighted) data to reproduce the model. For normal users, owning the model weights essentially means owning 100% of the model, you can inspect and study the weights in much the same way as the lab that produced the model can, you can modify the weights, and you can use and distribute them if the license allows you to