@jon_stokes I could be wrong here, but don’t all large consulting houses have enterprises licenses which prevent the model provider from training on their data (tokens).
If this is the case, how are they letting the LLMs into the chicken coop? What am I missing?
This is what happens when a man who has never opened an engineering textbook tries to design a national energy system on X. Every claim here is wrong.
Wind provides 30% of UK electricity, not 300% of UK energy needs. Electricity is only about 20% of total UK energy consumption. Wind provides roughly 5-6% of the UK's total primary energy. He is off by a factor of ~50.
A "single £10 billion battery plant" to store it all? UK peak electricity demand is ~45 GW. Storing 12 hours requires 540 GWh. At realistic costs that's tens of billions for half a day. Wind can be calm for days. You would need multiple days of storage. That is hundreds of billions, not ten. His number is off by at least a factor of 5-10.
"Done in 5 years." The UK has been building wind for 25+ years and has reached 30% of electricity. Not total energy. Electricity.
"Paid for within 10." Based on what? He provides no math because the math destroys his argument.
But fossil fuel companies "won't let it happen." The conspiracy is always the excuse when the physics does not cooperate.
As enterprises start to scale AI agents across the organisation, a new set of roles will begin to appear inside Engineering and Operations functions.
Their job will be to thoroughly document the context and knowledge that currently sits in people's heads. The workarounds, the dependencies, the edge cases, the approval flows - all the detail that isn't meticulously written down today, but is known within the organisation through osmosis and has grown over time.
Dashboards, PowerPoints and process maps aren't enough for agents to perform accurately and consistently. Detailed documentation at a granular level is required, and most enterprises don't have it in the form (or detail) an agent can use.
This is the less visible side of the agentic conversation. Most of the focus is on the models and the demos which look great, but far less attention is being paid to the underlying work that needs to happen for any of it to operate reliably inside a real business.
I am BEYOND excited to announce that after 1 year off, the best show in science & deep tech is BACK.
S3 returns this Saturday!
Bigger and better than before.
this is actually insane
> be tech guy in australia
> adopt cancer riddled rescue dog, months to live
> not_going_to_give_you_up.mp4
> pay $3,000 to sequence her tumor DNA
> feed it to ChatGPT and AlphaFold
> zero background in biology
> identify mutated proteins, match them to drug targets
> design a custom mRNA cancer vaccine from scratch
> genomics professor is “gobsmacked” that some puppy lover did this on his own
> need ethics approval to administer it
> red tape takes longer than designing the vaccine
> 3 months, finally approved
> drive 10 hours to get rosie her first injection
> tumor halves
> coat gets glossy again
> dog is alive and happy
> professor: “if we can do this for a dog, why aren’t we rolling this out to humans?”
one man with a chatbot, and $3,000 just outperformed the entire pharmaceutical discovery pipeline.
we are going to cure so many diseases.
I dont think people realize how good things are going to get
Trying to master Claude Cowork and autonomous skills, but the pace of these releases is wild. Anthropic is dropping features faster than I (or Sonnet 4.6) can track them! 😅
Why am I unable to download my full chat history from chatgpt? Is this a difficult feature to build? It would really help me better analyse responses
@sama
If the UK government maintained the same expenditure as in 2018-2019, which was £852.7 billion, and the revenue collected in 2023-2024, which was approximately £1.1 trillion, the government would have a £247.3 billion surplus.
Not sure if it is just me, but #ChatGPT web browsing just doesn't work. Takes forever, then tells me it can't access the website. Tried with over 20 websites and nothing...
learn that when determining how sustainable a solution is, make sure it can solve the most complex issue it could face further down the line and not just the problem you're solving right now.
As we know, these systems are only as good as the underline models. But at least Google gives you a range of facts and figures from multiple sources in order for you to make your own decisions.