407 karma · joined June 21, 2025
But then he changed his tune? Even on LLMs...
Market frictions cause the problem to be solved multiple times.
LLMs learn the solution patterns and apply it devaluing coming up with solutions in the first place.
People not being okay with having to share their improvements not being able to use the software is by design.
I don't get how you get from there to some sinister hostage taking situation.
Also everyone that contributes to the previous LGPL verison probably contributed under LGPL only, so it is now just one guy...
I can configure my Snapdragon plastic laptop such that the fan doesn't turn on, so the body being metal isn't a requirement for not turning on the fan...
The metal is more "luxury", though.
One gets back 50% if reaching the destination is delayed by more than 2h. Schedule the journey such that this is probable, making the journey 50% cheaper. Potentially with being able to define where one should be stuck waiting for the next train connected with sight seeing opportunities (such as the nice quarter near the Frankfurt main train station -- old ECB building!).
What you want instead is that the users just describe their problem, as unbiased as possible and with enough detail and then let the expert come up with an appropriate solution that solves the problem.
I try to do that as well when going to the doctor.
There is no doubt it is worth a lot in very remote destinations. But by definition there aren't many people there. Thus limited market size.
Launches for US military hardware is a market, yes.
Other countries are going to want to use/build up their own capabilities, so SpaceX won't have market access. Especially China.
Maybe then they'll notice that without Open Source training data it won't be able to solve the problem.
Currently it is probably just mostly quantity that does the trick w.r.t. training data. So e.g. spam the Internet with "product comparisons" featuring your product as the winner.
Currently it is in the "malicious compliance" phase.
LLM research will go back to (government funded) research labs with government funded supercomputers. All AI investment will need to be written off.
Running the LLMs the research generated will of course be possible, e.g. via AWS bedrock or alternatives. Initially there will be no "flat rate" subscriptions like currently (similar to early Internet), those will come once the prices are low enough further out. Running the LLMs is a low-margin business not justifying high multiples.
I concede that you'd have to trust that the id systems real identity => per site pseudo id mapping is not disclosed to anyone.
Listing was IMO a problem with minio as well, but maybe it is not that important because it seems to have succeeded anyway.
Now people holding actual cash will have problems.