I frequently push back on people being hair-trigger about calling things AI, but even I’ve gotta admit, that’s exactly what Claude code says if you ask it to do a security review and it finds something. I’ve seen this numerous times.
I can detect it pretty well, but that was just one example.
No person starts a summary that way, it's over-the-top and meaningless. I have seen AI do that many times when summarizing something related to security, though. Claude often says "CRITICAL:" or "CRITICAL VULNERABILITY:" or similar, especially when you jam the context window full of junk.
Was the golden boy for a while? What shifted? I don't even remember what he did "first" to get the status. Is it maybe just a case of familiarity breeding contempt?
It is starting to become clear to more and more people that Sam is a dyed in the wool True Believer in AGI. While it's obvious in hindsight that OpenAI would never have gotten anywhere if he wasn't, seeing it so starkly is really rubbing a lot of people the wrong way.
Well, in the world where AGI is created and it goes suboptimally, everybody gets turned into computronium and goes extinct, which is a prospect some are miffed about. And, in the world where it goes well, no decision of any consequence is made by a human being ever again, since the computer has planned every significant life event since before their birth. Free will in a very literal sense will have been erased. Sam being a true believer means he is not going to stop working until one of these worlds comes true. People who understand the stakes are understandably irked by him.
He is a pretty interesting case. According to the book "Empire of AI" about OpenAI, he lies constantly, even about things that are too trivial to matter. So it may be part of some compulsive behavior.
And when two people want different things from him, he "resolves" the conflict by agreeing with each of them separately, and then each assumes they got what they wanted, until they talk to the other person and find out that nothing was resolved.
Really not a person who is qualified to run a company, except the constant lying is good for fundraising and PR.
It's sort of two books combined into one: The first one is the story of OpenAI from the beginning, with all the drama explained with quotes from inside sources. This part was informative and interesting. It includes some details about Elon being convinced that Demis Hassabis is going to create an evil super-intelligence that will destroy humanity, because he once worked on a video game with an evil supervillain. I guess his brain was cooked much earlier than we thought.
The second one is a bunch of SJW hand-wringing about things that are only tangentially related, like indigenous Bolivians being oppressed by Spanish Conquistadors centuries ago. That part I don't care for as much.
Not a case, society call them sociopaths. Witch includes power struggle, manipulation and physiological abuse of the people around them.
Example, Sam Altman and OpenAI hoarding 40% of the RAM supply as unprocessed wafers stored in warehouses bought with magical bubble investors money in GPUs that don't exist yet and that they will not be able to install because there's not enough electricity to feed such botched tech, in data centers that are still to be built, with intention to punch the competence supply, and all the people of the planet in the process along two years (at least).
Yep the various -path adjectives get overused but in this case he's the real deal, something is really really off about him.
You can see it when he talks, he's clearly trying (very unconvincingly) to emulate normal human emotions like concern and empathy. He doesn't feel them.
People like that are capable of great evil and there's a part of our lizard brains that can sense it
Have worked in places where juniors had to lock devices when on prem; only authorized hardware in the rooms. Yet, the danger was from sloppy O6+ not the O1/GS6 who would (ready&abel) carry the water.
The is a serious problem with folk with power and authority and somehow no responsibility.
> The is a serious problem with folk with power and authority and somehow no responsibility.
Or perhaps the fundamental problem is with people in general - perhaps people without power and authority follow rules only because they don't have the power and authority to ignore them.
If the bubble is over all the built infrastructure would become cheaper to train on? So those open models would incenerate less? Maybe there is an increase of specialist models?
Like after dot-com the leftovers were cheap - for a time - and became valuable (again) later.
No, if the bubble ends the use of all that built infrastructure stops being subsidized by an industry-wide wampum system where money gets "invested" and "spent" by the same two parties.
There is this thing that happens in USA where RFPs are issued in such a way only one vendor could pass the mark - does that happen in UK? Reckon PwC has connections to make that happen
Probably depends on the department. I do grant and loan assessments for Innovate UK, and they have a rigorous and largely (+) transparent method for assessment which I would be happy to explain in detail. If we award money, it's accompanied by a monitoring officer (I do that as well) who is subject area expert with project management business experience. The MO meets the project every one or three months to review progress and approve payment of an installation of the grant or loan. We certainly wouldn't hand over £4M without good reason!
(+ Some of the detail of the scoring matrix is not as transparent as we would like, but Innovate UK does take feedback and tries to improve it).
It does to an extent but less so particularly from central government.
The tender is here [0], the approval process is usually pretty watertight. The contracts that go through this will have a paper trail. What you’ll likely find is that PWC has written a spec that meets the letter of the contract and they have delivered a site that meets the letter of their wording, which is what they’re good at. The fact that it didn’t actually solve the problem is inconsequential to PwC
> The fact that it didn’t actually solve the problem is inconsequential to PwC
You are mistaken. The fact it does not solve the problem is good for business, because follow up contracts to resolve any shortcomings will most likely also be awarded to PwC, since they are the only bidder to already have the in house expertise on this bespoke site...
I don't know of a department that does it as well as MoJ, though. Caveats exist around old private sector implemented systems like the prisons and probation databases etc, which even MoJ itself doesn't own the IP for. But everything made by civil servants or contractors at MoJ ends up published in that org unless there's a good reason not to.
Edit: FOI in principle allows you to request a cut of a git repo etc for a service, so you can impose annoyance upon departments that are less open.
I'm not quite as cynical as that to be honest. I do think it's abundantly clear that PwC and other major consultancies hire people who have experience in writing briefs to get major projects, and that they likely have an implicit preferential treatment becacuse they meet things like insurance requirements already without any overhead. It's also clear, allowed (and provable) that they will start with senior people who can make a good point and sell a good story, but will swap them out as soon as possible.
Social media companies are also regulated, but we are talking about whether social media companies should be liable for creating addictive content when porn has the same qualities of being easily available and free.
reply