Users complain of new "sycophancy" streak where ChatGPT thinks everything is brilliant.
See full article...
See full article...
As a British person, we would prefer Marvin. There's a reason D N Adams created and sent up the Sirius Cybernetics Corporation and its robots with relentlessly upbeat Real People Personalities, and then mentions that come the revolution they were indeed the first ones to be put up against a wall and shot.Were you expecting Marvin (the Paranoid Android) perhaps?
It is in fact yet another pseudo-scientific con at the production (not the research) level. One step up from the perpetual announcements of water-fuelled motors, one step below testing your DNA in the current state of knowledge.I genuinely wonder what they can actually do with models that remain largely black boxes other than add more background instructions to the initializing prompt to try and establish guardrails.
We know that even having the models 'reveal' their reasoning steps is hardly bulletproof and one of the whole points (both good and bad) about these things is that they're non-deterministic. There's no simple feature flag to toggle, no code to comment out.
It's also wild to me that they're asking businesses and investors to sign onto their platforms when you can have hugely impactful behavioral shifts just kinda...happen. This isn't the first time (remember the laziness thing?) and it won't be the last time that LLM 'productivity' is adversely impacted for unclear reasons.
It's like betting everything on a specific horse and rider in a race, but the horse is known to have spurious, uh...outbursts. Sometimes on the track.
Aldous Huxley beat him to it. In Brave New World there is limited automation so that the upper classes (the alphas and betas) will have things to do, managing the gammas and deltas. But the gammas and deltas are scientifically raised to be happy with their lot and enjoy their simple tasks. An elevator could be controlled by buttons, but instead a happy delta takes it up and down all day before going off to simple communal games with other deltas.I was hoping that somebody would chime in with a Douglas Adams reference. My first thought when I read the headline was that our future might really include things like elevators that sigh with satisfaction at having delivered us successfully to our floor. As usual, Douglas Adams was ahead of his time...
I am quite sure that had we had this in the early 1990s I would have known because of the superior literacy. (How does a CEO with a degree in journalism manage to be semi-literate?)On the plus side, lately I've been able to immediately tell when my boss sends me something straight from chatgpt becuase of its chipper tone and frequent use of emojis that are uncharastic of a gruff man in his late 50's.
Context helps. This was a manufacturing company. The CEO had risen without trace through relationships. Nothing to do with publishing.I've been in journalism since the '90s. At no point did a CEO anywhere I worked (nor publishers for privately held companies) have a journalism degree. They have MBAs and no understanding about how news works, because getting it right is a pittance against making more money.
I went to a conference for college newspaper editors where https://www.amazon.com/s?k=when+mbas+rule+the+newsroom&crid=57BQ07SZLMG7 was included in the goodie bag.
We fucking knew about this and actively chose to ignore it. It remains one of my only physical books.
Soft (in the head) bank? I wish I had $40G to invest that I could direct towards a virtual mistress.That attitude got them $40 Billion in investment from Softbank. Masayoshi Son (Softbank chairmain and CEO) has talked at length about how much he's enjoyed talking to Chat GPT late at night.