Send Link mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://mailman.anu.edu.au/mailman/listinfo/link
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Link digest..."
Today's Topics:
1. Re: AI search engines (Stephen Loosley)
2. O/t: Breakthrough Genomic Test Diagnoses Nearly Any Infection
(Stephen Loosley)
3. O/t: China's dominance in global shipbuilding strengthens
amid surging global-demand (Stephen Loosley)
----------------------------------------------------------------------
Message: 1
Date: Sat, 21 Dec 2024 12:51:01 +1100
From: Stephen Loosley <[email protected]>
To: [email protected]
Subject: [LINK] Re: AI search engines
Message-ID:
<sy5p282mb44098fef87721bc35b97095ac2...@sy5p282mb4409.ausp282.prod.outlook.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Thanks for this work Tony .. and perhaps of interest ..
OpenAI Skips o2 and Debuts New o3 ?Reasoning? Model
Reasoning models are supposed to fact-check themselves by producing a
step-by-step plan to find a correct answer.
By Thomas Maxwell Dec 20, 2024
https://gizmodo.com/openai-skips-o2-and-debuts-new-o3-reasoning-model-2000541796
OpenAI has unveiled Version o3, a new chain-of-thought ?reasoning? model
that the company claims is its most advanced yet.
The model is not yet available for general use, but safety researchers
can sign up for a preview starting today.
https://openai.com/index/early-access-for-safety-testing/
OpenAI and others hope that reasoning models will go a long way toward
solving the pernicious problem of chatbots frequently producing wrong
answers. Chatbots fundamentally do not ?think? like humans and different
techniques are needed to try and create the best simulacrum of a human
thought process.
When asked a question, reasoning models pause and consider related
prompts that could help produce an accurate answer.
For example, if you ask the o3 model, ?can habaneros be grown in the
Pacific Northwest,? the model might lay out a series of questions it
will research to come to a conclusion, such as ?where do habaneros
typically grow,? ?what are the ideal conditions for growing habaneros,?
and ?what type of climate does the Pacific Northwest have.?
Anyone who has used chatbots knows you sometimes have to prompt a
chatbot with additional follow-ups until it finally gets the right
result. Reasoning models are supposed to do this additional work for you.
o3 is the successor to o1, OpenAI?s first chain-of-thought reasoning
model. Reps said they decided to skip the ?o2? naming convention ?out of
respect? for the British telecommunications company, but it certainly
doesn?t hurt that it makes the product sound more advanced.
The company says the new model comes with the ability to adjust its
reasoning time. Users can choose low, medium, or high reasoning time;
the greater the compute, the better o3 is supposed to perform. OpenAI
says it will spend time ?red-teaming? the new model with researchers to
prevent it from producing potentially harmful responses (since again, it
is not a human and does not know right versus wrong).
Reasoning is the buzzword of the day in the field of generative AI, as
industry insiders believe it is the next unlock necessary to improve the
performance of large language models. More compute eventually does not
offer equivalent performance gains, so new techniques are needed.
Google DeepMind recently unveiled its own reasoning model called Gemini
Deep Research, which can take 5-10 minutes to generate a report that
analyzes many sources across the web in order to come to its findings.
https://gizmodo.com/google-releases-faster-gemini-2-0-with-deep-research-2000537349
OpenAI is confident in o3, and offers impressive benchmarks?it says that
in a Codeforcing testing, which measures coding ability, o3 got a score
of 2727. For context, a score of 2400 would put an engineer in the 99th
percentile of programmers. It gets a score of 96.7% on the 2024 American
Invitational Mathematics Exam, missing just one question.
We will have to see how the model holds up in real-world testing;
OpenAI?s recently released Sora still needs work. But optimists are
confident that the problem of accuracy is being solved. Still, tread
lightly relying using AI models for important work where accuracy is
necessary.
AI model companies like OpenAI and Perplexity are in a race to become
the next Google, collecting the world?s knowledge and helping users make
sense of it all.
They even have search products now that are meant to more directly
replicate Google with access to real-time web results.
All of these players seem to leapfrog one another with every passing
day, however. The feeling is somewhat reminiscent of the late ?90s when
there were a myriad of search engines to choose from?Google, Yahoo, and
AltaVista, Ask Jeeves, just to name a few, all hoovering up the
internet?s data and presenting it just with a different UX.
Most of them disappeared after one came along that was supremely better
than the rest?Google.
OpenAI clearly has a strong lead right now with hundreds of millions of
monthly active users and a partnership with Apple, but Google has
received a lot of plaudits recently for advancements in its Gemini models.
The Verge reports that the company is going to soon integrate Gemini
more deeply into its search interface.
--
------------------------------
Message: 2
Date: Sat, 21 Dec 2024 09:38:00 +0000
From: Stephen Loosley <[email protected]>
To: "[email protected]" <[email protected]>
Subject: [LINK] O/t: Breakthrough Genomic Test Diagnoses Nearly Any
Infection
Message-ID:
<sy5p282mb44094160c5421d09d6f2ab30c2...@sy5p282mb4409.ausp282.prod.outlook.com>
Content-Type: text/plain; charset="utf-8"
Breakthrough Genomic Test Can Diagnose Nearly Any Infection
Routine mNGS virus detection in clinical and public health laboratories possible
Granted breakthrough device designation by the US Food and Drug Administration
(FDA) in August of 2023,
https://www.reddit.com/r/technology/comments/1gsw6zf/breakthrough_genomic_test_can_diagnose_nearly_any/
Vailhem
OP 1mo ago
Top 5% Poster
Laboratory validation of a clinical metagenomic next-generation sequencing
assay for respiratory virus detection and discovery - Nov 2024
https://www.nature.com/articles/s41467-024-51470-y
Abstract
Tools for rapid identification of novel and/or emerging viruses are urgently
needed for clinical diagnosis of unexplained infections and pandemic
preparedness.
Here we developed and clinically validated a largely automated metagenomic
next-generation sequencing (mNGS) assay for agnostic detection of respiratory
viral pathogens from upper respiratory swab and bronchoalveolar lavage samples
in <24?h.
The mNGS assay achieved mean limits of detection of 543 copies/mL, viral load
quantification with 100% linearity, and 93.6% sensitivity, 93.8% specificity,
and 93.7% accuracy compared to gold-standard clinical multiplex RT-PCR testing.
Performance increased to 97.9% overall predictive agreement after discrepancy
testing and clinical adjudication, which was superior to that of RT-PCR (95.0%
agreement).
To enable discovery of novel, sequence-divergent human viruses with pandemic
potential, de novo assembly and translated nucleotide algorithms were
incorporated into the automated SURPI+ computational pipeline used by the mNGS
assay for pathogen detection.
Using in silico analysis, we showed that after removal of all human viral
sequences from the reference database, 70 (100%) of 70 representative human
viral pathogens could still be identified based on homology to related animal
or plant viruses.
Our assay, which was granted breakthrough device designation from the US Food
and Drug Administration (FDA) in August of 2023, demonstrates the feasibility
of routine mNGS testing in clinical and public health laboratories, thus
facilitating a robust and rapid response to the next viral pandemic.
Upvote 28
------------------------------
Message: 3
Date: Sat, 21 Dec 2024 22:52:42 +1030
From: Stephen Loosley <[email protected]>
To: "link" <[email protected]>
Subject: [LINK] O/t: China's dominance in global shipbuilding
strengthens amid surging global-demand
Message-ID: <[email protected]>
Content-Type: text/plain; charset="UTF-8"
China?s dominance in global shipbuilding strengthens amid surging global demand
China?s shipyards are set to be the main beneficiaries of a wave of new orders
as their South Korean competitors adopt a more ?cautious? approach, according
to ING
By Ralph Jenningsin Hong Kong and Frank Chenin Shanghai Published: 5:00pm, 18
Dec 2024
https://www.scmp.com/economy/global-economy/article/3291320/chinas-dominance-global-shipbuilding-strengthens-amid-surging-global-demand
China, the world?s largest shipbuilder by market share, is expected to receive
a raft of new orders as the global industry enters a renaissance, analysts said.
Demand for new ships is expected to surge over the next few years, as shipping
companies have a ?pressing need? to replace ageing vessels and comply with new
environmental regulations, according to Min Joo Kang and Rico Luman, economists
at the financial firm ING.
And China is set to be the main beneficiary of this market upswing, with its
primary competitor South Korea likely to take a more ?cautious? approach by
focusing on profitable and reliable orders, the economists said in a research
note released on Monday.
?Over the past two years, we have seen some previously closed shipyards in
China reopen and start picking up orders again,? the analysis said.
?The largest fraction of the fleet eligible for replacement consists of bulk
carriers, which are predominantly built by Chinese shipyards.
?This could drive a stronger need for investment expansion in China.?
London and Shanghai-based industrial consultancy Hartland Shipping Services has
also predicted sustained growth for the entire sector, with the prospects
particularly bright for Chinese shipyards.
In October, market demand for new-build container ships reached its highest
level since the second quarter of 2021, the peak of the last wave of global
shipbuilding, the company said in a report published that month.
The need for new ships is being driven in part by the introduction of the UN
International Maritime Organisation?s Energy Efficiency Existing Ship Index and
Carbon Intensity Indicator in January 2023, which mandate that all ships
calculate their energy efficiency and carbon ratings to cut pollution,
according to ING.
Global shipbuilding boomed between 2002 and 2008 due to China?s rapid economic
growth and an expansion in global trade, but the industry suffered a downslide
following the global financial crisis in 2008.
China also emerged as a dominant player in the shipbuilding industry during
that period, with its builders churning out container ships, bulk carriers, oil
tankers and liquefied natural gas (LNG) carriers.
According to shipping market analysts Clarkson Research, Chinese shipyards
commanded an overwhelming 86 per cent of global orders for new ships as of
September, followed by South Korea with a 12 per cent share.
China accounted for 55 per cent of the global order backlog as of the end of
September, with South Korea making up a 26 per cent share, Clarkson Research
said.
China?s rise as a leading force in the global shipbuilding industry can be
attributed to its large-scale production capacity, strong workforce, investment
in new technology and government refund guarantees.
Labour in China?s shipbuilding industry costs 50 per cent less than in South
Korea or Japan, according to ING, and Chinese companies also benefit from
access to cheap steel.
At the same time, the Chinese government provides ?sovereign refund guarantees?
for certain classes of vessels, easing financial burdens on shipyards, ING said.
Though South Korea?s market share has fallen over the past four years, its
shipbuilders ?lead in efficiency?, ING said.
High-value ships such as LNG carriers are often sourced to Korean builders, it
noted.
Shipbuilding accounts for a greater share of total exports in South Korea than
in China or Japan, according to ING.
In China, most orders are placed domestically.
--
AHWC
------------------------------
Subject: Digest Footer
_______________________________________________
Link mailing list
[email protected]
https://mailman.anu.edu.au/mailman/listinfo/link
------------------------------
End of Link Digest, Vol 385, Issue 25
*************************************
