Iranโs attack on Israeli just now (below) with several ballistic missiles is Tehran re-writing the โequationโ of escalation, so that it will now connect Beirut to Tehran, meaning it claims it is responding to an Israeli strike on Beirut, which Israel says is a response to Hezbollah.
Iran has been trying to redraw the borders of deterrence and escalation over the last month the, clawing back what it wants and pushing for more in the Gulf
These are dangerous times
The growing risk of a collapse of the Atlantic Meridional Overturning Circulation system, of which the Gulf Stream is part, is nothing less than the number one long-term security threat to our way of life in Britain, Europe and the Western world in the era in which we live.
The consequences on our societies of an AMOC collapse would be simply devastating for Britain especially, beyond anything imaginable but a full blown super pandemic or nuclear war โ with scientists modelling temperatures dropping around 15.c and half of our arable land being lost.
This is just one of many climate catastrophes starring at us of the modelling and the observed data and is why it is why Labour has continued to place such importance on Net Zero and international climate talks despite the Greens and progressive activists now looking elsewhere post-October 7th and the Conservatives joining Reform in now campaigning against them.
It is also one of the most important areas our foreign policy has and must continue to diverge from Trumpโs America, an active opponent to progress on this agenda. It is also why despite the huge security, technological and industrial risks they pose we need an active and substantive dialogue here as far as is possible with China โ which Labour has pushed for.
Google has published a paper that might end the transformer era.
For the last 7 years, every major AI, ChatGPT, Claude, Gemini, has been built on the exact same architecture: The Transformer.
But Transformers have a fatal flaw.
To remember context, they have to process every single word against every other word. Itโs called quadratic complexity. As your prompt gets longer, the compute cost explodes.
The alternative is the old-school RNN (Recurrent Neural Network). RNNs are incredibly cheap and fast, but they have a fixed memory size. If you give them a long document, they get amnesia.
Until today.
Google researchers published Memory Caching: RNNs with Growing Memory.
And it fixes the biggest bottleneck in AI.
Instead of an RNN having a fixed, rigid memory that constantly overwrites itself, Google gave it a "save" button.
The technique allows the RNN to cache checkpoints of its hidden states as it reads.
The memory capacity of the RNN can now dynamically grow as the sequence gets longer.
They built four different variants, including sparse selective mechanisms where the AI actively chooses exactly which checkpoints matter most.
The results rewrite the rules of efficiency.
On long-context understanding and recall-intensive tasks, these new Memory-Cached RNNs closed the gap with Transformers.
They achieved competitive accuracy without the explosive, quadratic compute cost. It perfectly bridges the gap between the cheap efficiency of an RNN and the massive capability of a Transformer.
We have spent billions scaling Transformers because we thought they were the only way an AI could remember a long conversation.
But Google just proved we don't need to process the whole history every single time.
We just needed a smarter cache.
SAM ALTMAN HAS A NEW PROBLEM. ๐คฏ
Google just shrunk 31GB of AI memory down to 4GB.
The tool is called TurboVec.
It uses up to 16x less memory, searches faster than FAISS, runs fully offline, and works on a regular Mac.
No expensive GPU cluster.
No cloud dependency.
No compromise on speed.
โ 16x lower memory usage
โ Faster vector search
โ Works with LangChain & LlamaIndex
โ 100% open source
The race to build bigger AI models is loud.
The race to make them dramatically cheaper just got a lot more interesting.
Repo: https://t.co/08TFGtHL6K
@preetkailon From 472$ to 412$, 60$ in three days. Incredible, maybe tomorrow 80$ in 4 days, to 392$. At least 70$ I think to 402$. But, that would be a buy.
So sahen die Umfragewerte aus, bevor die Journaillen der Milliardรคre und Fossilbonzen, insbesondere die Springer-Presse, eine nie dagewesene Fake-News- Hetzkampagne gegen Habeck fuhren...