No doubt Fable can outperform on many coding tasks, but the way you phrase it really exaggerates the gap and I'd suggest that for 95% of tasks, most devs simply don't need Fable level, and in fact, despite mostly using Claude, I have found codex is generally quicker at most tasks and can be even better than Fable with certain languages.
This.
And maybe anecdotal, but my experience of people who have used melatonin is that they sleep very deeply, but for a shorter amount of time. But likely they will wake up after around 6/7 hours sleep, rather than the recommended 8 hours of sleep. And once woken up, it's very hard to go back to sleep.
Funnily enough, I ran across an article this morning titled "What you hate about AI is the capitalism." and I feel like you can basically apply that to all tech!
It’s probably naive to assume a different system would cure all our ills. But I do wonder - if humanity acted for the collective good instead of individual self-interest, how different would our technology be? And how different would we be because of it?
I agree - I think we really need to encourage reduction in the obvious negative habits and encourage people to take up more positive habits. eg.
- Use car less; walk, cycle, or use public transport more
- Eat meat & air-freighted food less; eat local/seasonal produce more
- Treat household items as disposable less; repair what breaks and buy for longevity more
I am sure there are plenty more, but how to incentivise these things is a hard problem. One exception being food, where you could easily shift agricultural subsidies away from intensive factory farming and towards local/seasonal food (details may be more complex, but it's do-able!)
Is it possible that these companies simply fork the existing protocols and use the technology without buying into the existing "crypto world"?
Yes it won't be quite so decentralised, but say a number of major banks all spin up a node for say a JPM asset trading blockchain, it becomes semi-decentralised, so they have some advantage of a using a more secure shared ledger, but they also retain more control and thus probably more acceptance within banking, as big players can keep a walled-garden of sorts.
Completely disagree that 99% of people will just do it.
A lot of people are already skeptical of the frontier labs - I moved over to Claude from OpenAI when they bent over for the US government. And I'll certainly move on again if Claude start asking for photo ID.
The TLDR is that the best setup is probably Mac Studio (128GB RAM) / MacBook (36GB) with Qwen 3.6 35B (3B active params), or Qwen 3.5 122B model (this one is slow though).
These models are still very capable with good hardware, but they do lack the deep reasoning of major models and require more precise prompting.
So unless you really need the privacy, or have a lot of excess cash, it is not recommended, as considering the price of major models, it's just extremely cost inefficient!
Since when did the Loop demonstrate it could beat subway systems?
A heavy rail subway line can transport 30,000+ people per hour compared to vegas loop which maxes out at 4500 people/hr. Plus every single Tesla in the tunnel requires a human driver, unlike trains.
Even IF Elon managed to create these larger Tesla vans, they won't reach those numbers on a single route.
Not a fair comparison... A model can ingest a countless number works in day and reproduce stylistic fingerprints on demand, at zero marginal cost. How are the people it learned from meant to compete with that?
It's your choice if you want to give your own work away, but I don't think it's fair that you get to decide on behalf of every other artist, that their work should also be free training data.
Do you want all musicians and artists to put their work behind paywalls? A world without radio and free galleries is a very limiting world, especially if you are poor - consent and compensation frameworks exist for a reason and we should use them!
That would be a dream, but cannot see it happening.
But totally agree with your theory- platforms should face genuine legal exposure for algorithmic harm to minors (as tobacco companies did for health harm).
Unfortunately, as we found out recently, Meta's lobbyists are a powerful force to contend with and I do not trust our governments to stand up to them.
Well yes, maybe a bit of both?
Under the current administration, the dairy lobby has moved from being defensive (protecting subsidies) to being offensive - leveraging Secretary Kennedy's love milk/dairy to expand their market share within federal health policy.
These PACs have a lot of money to throw around, so I am naturally a little suspicious. And I'll give him putting steak at the top of the new food pyramid , but having cheese literally at the top?? That's too much...