Member Info for moonparty

My take is that results have done little for the SP in the last 12-24 months, because any short-term price rise has been offloaded into by our persistent seller(s).

However, there are signs that that supply is becoming exhausted, so a positive update just might be a catalyst to a higher price this time, even if without funding/partnership details. I live in hope anyway.

Doesn't really mater if the million was a buy or sell, the price action strongly suggests the overhang is cleared.

I think the seller(s) might be out or nearly out at this stage:

- price has been slowly rising all week

- whilst there have been some delayed trades, they have not been as big

- you can sell very high volume online - currently 500,000, has been up to a million this week.

1,081,202 is a very exact number - possibly the dregs. Here's hoping, if so goes up a bit on Monday that may have been the end of the supply.

Can't read much into one result, but PFS is looking very strong here.

Seems several institutions are accumulating

Having regular first hand experience of an inbox of several hundred emails, all of which require a response, you have to prioritise. And, with due respect, I would prefer Scancells priorities to be on other matters.

They are usually pretty good at responding, so I would take this as a positive.

They are probably too busy, and there was nothing new on the slides. The only 'new' one was the one with redacted info about who they are in discussions with, and that was really just a prop to make the point that there is a lot of interest .

Actual bid is 9.99p for 100k, 9.6p for up to 600K on auto!

Suspect the official bid will move up soon

Balerno, the LLMs (Large Language Models) that ChatGPT and similar offerings use are pre-trained and closed, so not trainable in the traditional neural network way that you may be thinking of.

However, they are dynamic in that they will do an internet search for every request (there will be caching going on, but essentially this is true). So, if there is differing relevant information available on the internet then the result may be different.

Just to expand a bit on EE’s very valid points about AI, it will lead to bias if you ask it biased questions, which is not helpful. It becomes an echo chamber.

e.g. a question such as ‘why is scancell undervalued’ is biased and AI will find all the positive opinions on scancell because of the context you have given it.

A better question, ‘what is the value of scancell’, will probably give you the current AIM valuation and possibly the brokers target prices, but that's not interesting and doesn't give you the dopamine hit. The temptation is to keep refining the question until you get the answer you want.

AI is not actually intelligent, it is essentially the ability to understand and generate language combined with the ability to search the internet.

To use AI effectively you have to ask neutral questions, and also be aware that it is pointless to ask it stuff that doesn't have an answer. You will still get an answer, but it will be about as valid as asking your astrologer.

My first Scancell AGM, not much to add to what has been said. The board looked and acted with amazing confidence and I am extremely reassured that it's game on.

I had a very brief chat with Mr Diggles in the gents as he was on his way out. Very upbeat . "Hang in there" were his parting words.

I'm also planning on attending the AGM as I'm in that London for business. Should we all wear a carnation or something? 😄

I'm going out on a limb and predicting good news tomorrow.

No rationale except that it's my birthday and I feel I am overdue some positive karma from the Universe!

When I checked earlier you could sell up to 750K online, smaller volumes at higher than mid.

Looks like someone is mopping up all the sells and wants more, but with a limit price < 10p.

Https://pmc.ncbi.nlm.nih.gov/articles/PMC12144009/

Thanks Bermuda, I had missed that one.

It doesn't really make it clear why they are reducing the trial size, beyond "increased focus". I guess, as you say, they were running out of cash?

Thanks for the reply Bermuda, I can't provide a link as I can't find any rns or other source that states definitively why the patient numbers were reduced from the planned 75 to 16. I searched through RNS and the website and asked several different LLMs (AI) - the LLMs suggested what I posted earlier about patients being reassessed after 12 weeks, and also mentioned "recruitment difficulties".

As I said, happy to be corrected on this.

Got my maths wrong - actual ORR would be 24/12, so only 50%!

Was thinking more about this on my weekend run this afternoon.

It's difficult to make any accurate deductions without knowing how many patients were considered and subsequently rejected at 12 weeks. Evaxion haven't published that data, or at least I can't trace it down!

If they initially recruited 75 and rejected 59 at 12 weeks, then, IMO, the trial is worthless, as it is such a honed subset that you would expect fantastic results just on CI. It might be useful for validating their AI, but as a commercially viable drug it has no value.

So let's be generous and assume they actually recruited half the 75, so 32, and let's again be generous and assume that of the 16 rejected, half had passed away and half had progressed beyond eligibility. The only ‘fair’ way to compare to scib1/iscib1+ is to include those that progressed, so an extra 8 patients. That would take the actual ORR to 12/20 (60%) rather than 12/16 (75%).

I really can't see this as any meaningful competition to iscib1+