Are the people that had achieved degrees and other certification in a fraudulent fashion going to be stripped of those and potentially fired or expelled?

This is not an isolated incident and there have been many stories of people achieving via fraudulent means and consequently we are all worse off for it. IMO they need to be named and shamed. Stripped of whatever they gained by cheating.

Many years ago we had a new guy with a Masters in networking. Didn't know what are the network ranges for class A, B or C were. Couldn't say what the difference is between TCP & UDP. Had no idea about subnetting etc etc. His masters seemed more related to social networking.

Ultimately an individual that gets certified or qualified in a fraudulent fashion is genuinely cheating themselves. They always run the risk of ventually getting found out and sometimes in very embarrassing fashion.

P.S. - we trained LLMs on human data...is it a wonder why they are sometimes dishonest?

I was wondering what would make people get back to other browsers...

Considering everyone is abusing on chromium can you also make sure to inject your enshitification intravenously in a way it cannot be removed?

Thanks Google!

They want people with AI skills but the whole point of AI is to eliminate the meanial and the repetitive "do key work" to do the high value work.

In other words employers still need professionals and Subject Matters Experts to tell the AI how to do things, what a good job looks like, how to check it, where the AI is shit, what to change etc etc.

I used AI every day. I program with AI every day. I used the same agents everyone can access but I get better results because you need to know what good is, you need to know how to define it, check it, explain it, target it, what research is relevant and what is an AI hallucination and so on.

Telling people how to use the AI is easy. Telling them to be engineers is hard/impossible.

When your AI solution runs out of context or tells you "Yes, I'm sorry that was the wrong DB table" you still need that human expert.

Microsoft word did not make printers extinct. Excellent did not eliminate accountants. CNC machines did not eliminate carpenters etc.

If anyone asks you to have AI skills on a job interview ask them what training they provide to do the role they are advertising. Claude doesn't know about their specific company, their process, their policies, their documents pack and that annoying quirk they have in their systems because #Reasons - that has to be covered by training.

The best employers look for the right people with enough knowledge to be able to absorb the training at a reasonable period of time and start running.

As the maximalists say; 1 Bitcoin is still worth 1 Bitcoin.

See you guys at sub $20k -Apologies in advance.

Dd iyou forget the 30 years or so after joining the EU that the UK flourished?

The UK suffered after the dot-com bust and to compensate they borrowed and borrowed. Suffered heavily from the 2008 financial crisis and it borrowed and borrowed some more...

Maybe the problem was with those that were pushing solutions that did not treat the problem?

The UK today has such severe problems that even a Labour government is forced to take austerity measures.

There's a lot of pomp, arrogance and bravado amongst some so called ruling class elites in the UK...nut there's nothing to base this on. British exceptionalism is dead.

This is the wonderful about democracy. You don't have to wait for 20 years if you change your mind. Most country let you do that every 4-5 years.

I like the sound of that. Imagine if Anthropic moved their HQ to the EU where the bureaucracy is so complicated even if the president wanted to do something there would be 17 objections and 23 years of legal battles ahead.

They'd be arguing if Fable 8.3 is allowed to teleport people by then.

Jokes aside, instead of bending the knee to Orange Jesus Anthropic should seek to migrate away from the US. Just assume now that he will rig the elections somehow to win a third term...because if you're counting on him disappearing I just want you to think about what happened last time...

That's the sort of commitment we need. Years of documenting and investigating all of Epstein's friends, accomplices, business partners and those in government that are obstructing said investigations.

Apparently there's many thousands of mentions of current US government officials in those files in very compromising fashion.

Maybe the government can get Anthropic's Fable 5 to do a thorough analysis and sign all the files, their dates, counts, mentions of people and interactions with all metadata into a publicly verifiable library.

After all, the current POTUS is a huge proponent of transparency and honesty. Some say the most transparent and honest president ever, in the history of all mankind. All the people that work for him say so.

GitHub commits up x14 or something like that...AI is accelerating development and we'll only slow down if we have a consequent emergency.

...but more to the point AI is helping find and fix more bugs and security issues than ever before. This is a good thing.

What about all that AI all those big tech companies are investing in?

I appreciate this will be unpopular and brings it's own set of difficulties but we could use AI to filter images locally. Like a smarter Net Nanny. Yes some will get around it but most will not.

I'm not advocating for it but it's a possibility.

Not so clear cut. There are a lot more uses for the power output per square meter from nuclear submarines, satellites and space stations to replacing clean energy in countries where solar exposure is low and heat generation is a bonus.

Also, power source divercity adds to grid resilience.

I'm all for renewables but there are genuine challenges in storage and when nature is not playing ball.

...to help Microsoft. A worthy charity?

You are making silly assumptions about what I know...in any case every single emergence of a new threat has been met with a new defence, strategy, mitigation etc. When the next new scary thing comes it'll be overcome too. I point back to the many generations of humans that made it this far.

Maybe this is what the church missed. Evolution.

Sometimes Opus will cede that it doesn't do certain things very well and "level with you" - it is IMO far more balanced in it's expectation settings than other models...it also depends on how you refine it with skills and CLAUDE.md file. Do you wish it to declare assumptions? Silent errors? -It can do that and I find it helpful.

In fact you can Ralph loops it for X attempts to achieve Y goal because it does fail and sometimes fails to assess it's own capabilities to achieve the goal you've set out.

Much depends on the user. If you follow the 4D Framework for AI fluency, build a process with well defined skills then you will encounter less "moments of honesty" that it cannot achieve or unlikely to achieve...but it'll still happen on occasion.
https://www-cdn.anthropic.com/...

Insensitive clods.