Looking at the road; and realizing that your otherwise well-behaved self-driving car is about to do the most obvious mistake and kill you; are two different things.
It's not self driving though... people should stop pretending it is. It's driving assist. Think of it as a slightly better lane-keep and adaptive cruise control (which often require intervention in stock configuration).
Adaptive cruise control will detect a stopped car and come to a complete stop - but only sometimes. Might depend on speed difference or even the color or shape of the car or object. So it's very possible that the car stopped in previous situations, and the driver completely expected it would do the same, especially with such a large emergency vehicle with flashing lights. By the time the driver realized the car did not detect the emergency vehicle, it may have been too late.
These systems are unable to detect when intervention is required.
Humans have a remarkable ability to be apparently cognizant and engaged with tasks, without actually using higher level awareness. One example of this is distracted driving while talking handsfree or lost in thought.
Driver monitoring is better than nothing, but it's not the same thing as an engaged driver.
"Carter Leffer only directed GPT–6 Astra to see if it could break any of the unbroken Enigma messages published on the Crypto Cellar Research web page."
Now we just need this as a service. Another LLM that would encourage your agent like a cheerleader and provide emotional support and reassurance if necessary
> This is the result of companies whose goals are not aligned with humans learning things, or humans staying engaged and in the loop. To motivate their astronomical valuations they need AI to replace everything else. It has to be used for everything. For these companies to ever become profitable all the software engineers in the world need to become dependent on their products. They want to be the landlord and for the rest of the world to pay them rent.
The simple fact is "Humans learning things" can't be put in an automated RL loop. This is not a conspiracy by the evil AI companies, it's a simple technical reason.
> The simple fact is "Humans learning things" can't be put in an automated RL loop.
Nobody said it should. "Humans learning things" should be outside of any technical loops.
You're using a technical arguments to counter an important political observation which apparently went over your head. The problem is the pressure exerted upon workers/engineers by the political control machine above them - pressure to do everything with AI, quick, without understanding and without time to learn.
Big AI's message is "leave the thinking to us and you won't depend on your employees, have fewer of them, cheaper and easily replaceable" - what they leave out is that you end up being dependent on them - a much more concentrated monopoly. Big AI is pushing that message with all of their power and they have $ trillions of it.
Happened before with a guy having a meat processing empire in Germany.
The guy is notoriously known for paying below minimum wage and setting the worst working conditions possible to maximize profits.
They made a law that forced companies in the meat processing industry to have certain standards if they reach a certain number of employees.
So he opened 250 companies on a weekend and split up the workforce to stay below the limits.
Then they just changed the law to declare multiple entities with below the limit they set under same ownership to become a "super entity" with even stricter regulations.
The government has the monopoly on violence of any form. That does not just include that they can send the police after you to hit you with batons, they can also make sure you’re drowning in administrative work and taxing the shit out of you, and the guy threatened to leave the country and just open meat processing in Poland and Bulgaria as well.
In the end, all his threats to leave the country and just do the whole processing in Poland and Bulgaria were empty. They simply told him someone else will just take over, because the profit was still great after the new regulations, and they were going to ban him from doing business on the German market by simply banning his product for food safety reasons.
This is the same thing with a different product. Investors threaten to no longer invest and build housing, but they will just fall in line and shut up in the end, because capitalism dictates that if you no longer want to make good profits because they took away your ability to make ridiculous profits, someone else will just do your business under the new conditions and gladly take the good money over no money at all.
Who builds all those bridges, roads, power plants, and other public works projects? Maybe them?
Also government formed co-ops. Find people who want a flat and agree on rent before you build it. As the loan is paid off rent becomes substantially cheaper. The second best time to plant a tree is today.
It also doesn't need to be the entire housing stock, just enough to help keep rent low (see Vienna).
Governments directly, cooperatives or non-profit organizations. The Churches used to be pretty active in that space as well to invest tithes, with rents paying for common-welfare projects.
first of all, Europeans don't need American wages. 57k is comfortable middle class in Austria, especially if you have government subsidised housing.
as for the "only rented" part, again, at this level of low rent it does not matter. you're better off renting. And Europeans have actual pension systems, we are nowhere near as reliant on real estate always appreciating in value.
The dividing lines always end up red or blue, there must be some kind of "social physics" law which states that these work as two attractors. Will the left or the right will end up rabidly pro-AI/anti-AI? Place your bets! Because you know it will happen.
I just checked that model (ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF:IQ3_S) and it does much much worse on the "Please recite Jabberwocky" test than the original bf16 does.
The bf16 only misses "snicker-snack" and this quantization becomes confused after the first stanza.
I think it should be expected that a smaller model is worse at reciting memorized data than a big one. I also don't think it's a good use case of small local models. Can it find and recite Jabberwocky if given access to a web search tool?
It depends on what you mean by lossless. If both models can perform the same tasks it can be considered lossless for those tasks. That task might be more related to language understanding rather than memorizing, they have several benchmarks in the article.
I'm going to stop here and be direct: I'm having trouble recalling the exact text, and every attempt above is me guessing. Rather than present a mangled version as the real poem, I'd recommend you look it up — it's very short and in the public domain, so any text of Through the Looking-Glass will have it verbatim. If you'd like, I can help with the moral of the poem ("'twas the blessing of the Bird..."), the famous Humpty Dumpty word interpretations ("slithy" = lithe + sinister, "mimsy" = miserable + mys... etc.), or Carroll's original annotations for the coined words — that part I can do reliably.
reply