They train. Deepseek doesn't have a Zero Data Retention policy, but for many personal tasks I'm (not parent) ok with them having this data if it means better, cheaper models for everyone on the future.
Different story with $DAYJOB where ZDR matters, but the beauty of open models is that you can find a provider hosting them in your preferred terms.
> And the water used must be recycled and self sustaining. If you need a million gallons fine but that’s all we can’t have data centers sucking up all the water.
They can have a closed loop cooling system, but it costs more energy to run and would eat into their profits. so they use evaporative cooling, cheaper, but consumes entire towns worth of water. If this is pumped from the ground, round water levels dry up in all around and peoples wells dry up.
> 6. 200 OK, no answer
Reasoning models sometimes put everything in the reasoning field and hand back content: null, finish_reason: "stop". 345 completion tokens, HTTP 200, nothing to show the user.
A 200 tells you the request was served, not that there's an answer in it. No content and no tool call is a failure, throw and retry.
I ran into this so much I had to modify my harness to handle this.
> Do we use user feedback and de-identified data to improve ChatGPT and Codex in a holistic way? Yes.
This says that they trained on user sessions. The de-identification here I believe refers to removing PII, which doesn't matter here because the issue at hand is the content of the researcher's session where they likely discussed their approach tackling the Navier Stokes problem.
> Did any human or agent look at user data as part of the Navier Stokes effort? No.
If they trained the model on Lavent's chat sessions (PII removed or not), then this statement is meaningless as the model weights already contain that information.Given it's a new yet unreleased internal model, it is likely a 10+ trillion parameters (Astra is rumored to be 10 trillion), so the model can retain a lot more detail/info from training data.
Why is he leading the with the irrelevant part first ?
> And so does every LLM company.
Nope, not for enterprise users.No enterprise customers would use it if all of their internal business plans / trade secrets would end up in the model weights of the next OpenAI model. Imagine your competitor asking chatGPT a question and the model spitting out your business plan. These models can retain very specific fine grained data. I remember there were examples of them reproducing sections of their training data verbatim.
> Researchers had found signs of CKDu outside Central America, including in Sri Lanka, India, and Nepal. But it had never been documented in the United States.
I searched a bit more and found this interesting bit from an article talking about a similar epidemic in Sri Lanka.
> An analysis showed that trace metal ions such as calcium and magnesium that naturally occur in the hard water reacted with glyphosate. This meant that, instead of breaking down rapidly, glyphosate was able to persist in the environment for up to seven years in water and twenty-two years in soil.
> This meant that, instead of breaking down rapidly, glyphosate was able to persist in the environment for up to seven years in water and twenty-two years in soil.
Bulllshiiiitttt. There are no reputable studies that show anything like this.
The tl;dr - even if there were reputable studies, what do the results mean or imply exactly?
There is a study [1] which was via information from the link from the parent comment. Interpretation of results though is the issue. From their method of study, to myself it's clear that storing heavy mineralised ground water in a used Glyphosate spray tank or product drum, is a not a very good idea given the results of the study. It is though a different story for the groundwater given that degradation in soil [2] [3] is measured in half lives and depends on many factors. As for detection - there may well be some complex of Glyphosate that has resisted breaking down a couple decades later, but one would expect the amount would probably be far less that the allowable or typical amount of Glyphosate found in food from crops that have been sprayed with it prior to harvest.
The other question is what breakdown complexes are they measuring for detecting Glyphosate. For years roundup was though to be still be at higher levels of use than it should have been after it was phased out in some European countries ... based on levels detected in rivers. [4] [5] and the preprint via [6] indicate that some detergents were to blame for the higher detection level.
And to add a bit more context: there's a whole Glyphosate Griefing industry now that invents bullshit claims.
Pretty much every time you see something like "glyphosate reacts with kidneys resulting in hard water staying around for 111 years in dry soil" - you can be sure that it's this.
Half life [1] in biological systems isn't quite the same as radioactive decay. Essentially what is being relayed is that after a couple of months generally they are, by measuring the product directly or the secondary and tertiary by products, deducing a time when only half of the active product is left.
Farming systems that use herbicides, are going to cycle though every few years of using Glyphosate, so there's little point IMO of creating a special study that looks into Glyphosate residues over more than a few years where they are monitoring the degree of fractional amount that could remain - there are too many variations of soil types, rainfall and temperature to be greatly useful. Any results of how long it might persist would more likely come from real life examples - areas where its use has been phased out, but the difficulty most (if not all) would not have an accurate starting point of just how much Glyphosate was applied in total, time period and weather conditions to extrapolate carefully.
Detection is different to a noteworthy amount of product that is bound to clay or other particles which has prohibited biological degradation. Many who conduct studies in complex systems are usually focused on just a small number of instances they feel might encompass most of the issue they believe to be at heart ... to the point of dismissing local / on the ground knowledge.
Yes the study used fish which can be useful ... however they tried to simulate using the raw product. It would have been better had they had some water from a month after a paddock was sprayed and used surface or underground flow. What it actually means, for example to those like myself who use it regularly, is there is certainly evidence we have to be doubly careful to ensure no or as little as possible over spray of waterways occurs which contain aquatic life.
A better study might be what results might be when groundwater that has a noteworthy amount, is then heated with bound Glyphosate and it's by products, to simulate cooking and boiling of the water.
>"Nothing to do with soil persistance."
Do you think that if soil was measured for a study it was watered only with distilled water? The purpose of mentioning was to suggest that there's an uncertainty as to what may have been measured.
> Half life [1] in biological systems isn't quite the same as radioactive decay.
It typically is shorter, as real systems are often hyper-exponential. And if you're talking about persistance in isolated clays, then it's by definition is not a bilogical system. A sealed glass vial is technically a type of clay, and glyphosate can survive in sealed vials for decades.
There's a thing about glyphosate: it's the safest pesticide. It's non-toxic to mammals, is not carcinogenic, non persistent, and is highly useful. And this somehow triggers all the cranks because it apparently MUST be hiding something.
And glyphosate has been _extensively_ studied by legitimate scientists.
So at this point, pretty much all articles about glyphosate's harm are pseudoscience and/or transparent grifting.
Clays form a good proportion in some soil profiles in some parts of the world and very much part of the biological system even if it isn't that apparent.
Glyphosate and the by products are certainly safer than many registered herbicides currently available to much of the western world.
The wetter in the formulation of the official Roundup has certainly changed - what I used in the 1990's a few drops of the diluted ready to spray mix accidentally making skin contact would give me a bad headache even if I noticed it right away and washed it off ... when I used the official brand again (for a time, mostly 00's) they priced Roundup out of my farming budget) in the late 10's I noted that I was able to easily wash off any splash back without resulting in a bad headache.
I have been long following the twists and turns that certain big companies, one of which used to own the patent on Glyphosate until 1999 when they let it expire, with a plan to simply create a very large plant in iirc South America which would chase off the competition that have sought to produce Glyphosate for export usage. Unfortunately they put the price up (3x retail here in Australia) due to oil prices costs during the 00's and then China stepped up. I also know they have in place $$$$ BS products to suit their new patented seed which is herbicide ready which when I look over application requirements, complete 100% leaf coverage to work properly implies it is no where near as easy to use as regular roundup.
To be clear I really value Glyphosate as a herbicide as for agriculture it saves fuel, equipment wear and time when used properly. Even so since many are pointing fingers, it is important not to just gloss over any problems that using it might cause, even though I am quite sure there are worse herbicides to be had coming into use having observed this in my immediate locale.
> Vibe (Enterprise): customers are opted out of training by default
So they made it opt in for enterprise (how it should be), but intentionally made opt out for regular user.Basically saying "screw you: to regular users.
Agree, that it is an ad. Also, unless the full session (in a resume-able format, along all session artifact such as intermediate research/docs used to produce the final output) is also included in the commit, the session id is useless.
I presume this is Anthropic trying to their "usage" stats before their IPO.
> Anyone who did and could do so turned it off immediately
I know I’m not alone as someone who’s bothered by the postscript, but too lazy to do anything but delete it every time. There’s even an entire (well part of) song about it: https://www.youtube.com/watch?v=EbdeVhPAbms
Out of curiosity, do these Claude commit messages fall into the same category of "it's annoying but not worth the effort to turn it off"? If yes/no, why?
Most people's identities and self worth are tied up in the material things they buy rather than actual self actualization and personal axxomplishments.
Even as a lifelong Apple user and longtime iPhone user, that sig irritated the shit out of me and was instantly disabled whenever I set up a new device.
Also, and this may come as a shock to you and a number of other HN users: I never considered it a "point of pride." I switched because Android was a privacy dumpster fire comparatively and battery life was far better.
I also can't remember a discussion in person with folks where someone seemed "proud" to use an iPhone. If anything, the Android crowd loves to sneer at iPhone users and look down on them as idiots who need a "simpleton" phone.
I'm no simpleton. I ran custom roms and so on. Had F-droid installed, and so on. I do miss the customization, but I also look back and realize I wasted a massive amount of time on installing roms, backups, and the customizations themselves. Probably weeks of my life if not more. My iPhone? I don't. It's a phone. That runs apps. Do I wish I could install whatever I want, particularly given how secure apple's sandboxing is, and thus in theory how safe it should be? sure. but at the end of the day I want to live my life.
Now compare that to why people don’t want Claude in their commit messages or PR descriptions.
Clearly there’s no pride in using agentic coding (if anything, maybe it’s the opposite of pride..). I don’t think most developers want to publicize their use of agentic. But I could be wrong.
Do you know if they retain your prompts or use it for training?
reply