The model is not "open source", but it is an open weights model.
You can download it from the link given here at the top and you can run it on your own hardware, with whichever open-source harness you prefer, without having to worry about token cost or about subscription limits or about any future degradation in performance that you cannot control.
The recent history has demonstrated that such risks are very significant.
Being open weights is important for anyone who wants to use an LLM. Being open source is important only for a subset of those, who have the will, the knowledge and the means to train a model from its training data.
Having access to the training data used by a model would be very nice, but the reality is that for a normal LLM user it is very beneficial to use an open-weights model with an open-source harness, but it would be much harder to exploit the advantage of having access to all the information about how the LLM has been created.
For me open source means that the entire training data is open sourced as well as the code used for training it otherwise it's open weight. You can run it where you like but it's a black box. Nomic's models are good example of opensource.
Even with all training data provided, won't it still be a black box? Unless one trains it exactly the same, in the exact same order for each piece of data, potentially requiring the exact same hardware with specific optimizations disabled due to race conditions, etc., the final weights will be different, and so knowing if the original weights actually contain anything extra still leaves any released weights as a black box, no? There isn't an equivalent of reproducible builds for LLM weights, even if all of this was provided, right?
This is the kind of scenario that is served better by Go/C-style error values than exceptions. Error values facilitate and encourage you to log what you were doing at the precise point when an error occurs. Doing the same with exceptions idiomatically often requires an exception hierarchy or copious amounts of separate try/catches.
The difference really becomes apparent when trying to debug a customer's problem at 3am (IME).
Your stack trace tells you where in the code the error occurred, but doesn't tell you what it was doing with what data. For that you need to pass context for the error up the chain of calls, adding to it as you go up. Exceptions are not a great way of doing it as you only have the local context, which isn't a great help when you're catching N levels up.
And if you're not catching N levels up but catching at each level, then you are emulating error values but with try/catch blocks.
That is great for you as a developer. As a sysadmin supporting other people's crap, stack traces and heap dumps are useless beyond forwarding them to the vendor.
The article implies that this new one uses an IR image rather than just IR point sources, so can detect the shape of the aircraft among a cloud of decoy flares.
Pretty simple and probably quite effective if true.
> The newest generation of the FIM-92 Stinger uses a dual IR and UV seeker head, which allows for a redundant tracking solution, effectively negating the effectiveness of modern decoy flares (according to the U.S. Department of Defense).
This stuff is a constant back-and-forth of tech improvements.
"The Complete Memoirs of George Sherston" by Siegfried Sassoon. (Ignore the title, it's actually his autobiography, and you could probably skip the first book in the trilogy).
"Goodbye to all that" by Robert Graves.
Two of the best writers in the English language recounting their times in the trenches.
I've often thought world leaders, upon election/selection, should get a free few orbits of the earth, to give them some perspective on the job they're about to undertake. Maybe offer the first one on Artemis II, a deferred one for the current US administration?
James May of Top Gear has flown with a U2 spy plane once [0][1]. When they reached to the edge of space, May said "If everybody could do that once, it would completely change the face of global politics, religion, education, everything".
I can't agree more.
Another thing I believe needs to be watched periodically is Pale Blue Dot [2].
I think you overestimate the effect that would have on the kind of people that most need that sort of humility.
Look at what happened with William Shatner and Jeff Bezos when they came back from space. Shatner started to say something about what an impactful experience it was, but Bezos cut him off and was like “Woo! Partay!” and switched his attention to a magnum of champagne.
I met someone a couple years ago who was a U2 pilot (which are still in active service). He'd flown F-16s until he reached the point in the promotion ladder where flying stopped, then switched to U2s to keep being a pilot. After hitting 20 years, he was taking his retirement and training to fly Grumman S-2Ts with CAL FIRE.
Very down-to-earth guy who knew what he wanted and made his choices. Didn't at all seem like the sort to find edge-of-the-atmosphere flying a mystical experience.
Jeff went up two flights earlier, in July 2021 on NS-16. Shatner was on NS-18 in October.
I don't know if it's a thing that wears off, if Bezos was just in business-mode the entire time, or just didn't want someone monologuing right after getting back.
>I've often thought world leaders, upon election/selection, should get a free few orbits of the earth, to give them some perspective on the job they're about to undertake.
Perhaps, but they should also get a few free orbits of the Earth *after* their term ends, on a launch system built by whichever contractor has given the most "campaign donations" to politicians. Surely they'll trust it to be safe, right?
I would also say give them a year of free vacations in various places. Say a maximum security prison in general population, any type of dark camps, hospitals, mental institutions and care homes.
Give them the rest and recreation they need in these wonderful places.
"Houston, this is Golden One. I'm looking down on the big, beautiful, blue world. They love me down there. They all love me. I'm the greatest astronaut ever in the history of mankind. No one has ever orbited like this before."
I mean, we can probably predict what will happen based on existing data.
"I've seen things up there that are huge, absolutely huge. And let me tell you, astronauts, they came up to me, they were crying, big men crying. Earth, it's a good name, but it's not big enough, not grand enough. So, I'm thinking we rename it. How about 'The Trump Sphere'? It's got a nice ring to it, doesn't it? And let me tell you, nobody would argue with that name!"
reply