John Coltrane would have turned 100 on September 23, 2026, and the jazz reissue industry was never going to let that milestone pass quietly. That is hardly a complaint. We have spent a fair amount of 2026 listening to Coltrane again at eCoustics, including Craft Recordings’ excellent new Original Jazz Classics edition of his 1957 Coltrane debut.

We also recently reviewed the all-analog mono reissue of A Love Supreme, another reminder that the strongest parts of Coltrane’s catalog continue to reward both new listeners and those of us who have bought some of these records more than once.

Rhino is taking a different approach with John Coltrane 100: Selections From 1951-1967, a six-LP retrospective arriving October 2 for $149.99. Rather than concentrating on one album or one label period, Rhino has licensed recordings from Prestige, Blue Note, Atlantic, Columbia, and Impulse! to follow Coltrane from his early work as a sideman through the increasingly adventurous music of his final years. That makes this far more useful than another greatest-hits collection.

Related Reviews:

Five Labels, One Very Short Career

One of the difficulties with building a Coltrane collection is that his recorded history is scattered across multiple catalogs. The early years intersect with Dizzy Gillespie, Miles Davis, Thelonious Monk, Paul Chambers, and Art Blakey before Coltrane’s recordings as a leader begin to dominate the story. From there, the collection moves through Blue Train, Giant Steps, My Favorite Things, the Impulse! period, collaborations with Duke Ellington and Johnny Hartman, live recordings, A Love Supreme, and finally material from Stellar Regions.

Rhino has sequenced the box chronologically, which is important because Coltrane’s musical evolution happened at a pace that still feels almost ridiculous. The first LP includes early recordings such as the Dizzy Gillespie Sextet’s “We Love To Boogie,” an alternate take of “Two Bass Hit” with the Miles Davis Quintet, Paul Chambers’ “John Paul Jones,” and Miles’ “If I Were A Bell.” By the second record, “Blue Train” has arrived. The middle of the collection moves through “Giant Steps,” “Naima,” “My Favorite Things,” “Equinox,” and a live version of “India,” before the final side reaches “My One And Only Love,” a live “Alabama,” “A Love Supreme, Pt. 1 – Acknowledgement,” and “Stellar Regions.”

The value here is not merely having those tracks in one box. It is hearing the changes in Coltrane’s playing, compositions, ensembles, and priorities unfold in sequence, something that is much harder to appreciate when the albums are purchased piecemeal and listened to without much historical context.

There Are 37 Selections, Not 23

There is one odd discrepancy in Rhino’s own announcement. The company describes Coltrane 100 as having a 23-song track list, but the published six-LP sequence contains 37 selections across 12 sides. We counted them, and the published side timings add up to approximately 239 minutes, or just under four hours of music.

For anyone trying to determine exactly how much material is included, 37 selections is the number that actually matches Rhino’s own track listing. It is a strange error in the promotional copy, especially for a box built around a centennial celebration where collectors are going to look rather closely at what they are getting.

Technical Details

The six records are being pressed on 180-gram vinyl at Optimal Media in Germany, with mastering handled by Bernie Grundman. The set is packaged in a slipcase and includes a 12-page booklet featuring a new essay by Coltrane scholar Ben Ratliff. Rhino also says copies purchased through Rhino.com and participating independent retailers may include a limited 12 x 12 typography poster.

Those are all encouraging details, but there are still some important unanswered questions for audiophile buyers. Rhino does not describe the box as AAA, does not identify the source used for every selection, does not state whether the lacquers were cut directly from analog tape or from digital files, and does not specify the mono/stereo configuration of every track.

None of that means the records were poorly sourced or will not sound excellent. Bernie Grundman’s involvement carries considerable weight, but “remastered by Bernie Grundman” is not the same thing as documenting the entire mastering chain.

That matters because Rhino has been much more explicit when it wants to promote an all-analog release. Its John Coltrane: 1960-1964 Mono Rhino High Fidelity six-LP box was advertised as AAA, cut by Kevin Gray from the original mono master tapes, and pressed at Optimal.

Craft has been equally clear with its new OJC edition of Coltrane, while Blue Note specifies the analog mastering chain for its Tone Poet edition of Blue Train. Until Rhino provides more information, Coltrane 100 should not be described as an all-analog release.

Is $150 a Good Deal?

For six 180-gram records, the price works out to almost exactly $25 per LP, which is hardly outrageous in 2026, particularly for a box involving five different label catalogs, Bernie Grundman mastering, Optimal pressings, and nearly four hours of music. The more useful comparison is what happens if you start buying major Coltrane albums individually.

Current audiophile-oriented editions of Coltrane, Blue Train, Giant Steps, My Favorite Things, and A Love Supreme generally sell for around $39 each. Buying those five alone pushes the total close to $200, and you would still be missing the Miles Davis, Thelonious Monk, Dizzy Gillespie, Duke Ellington, Johnny Hartman, live Village Vanguard, and later Impulse! material represented here.

The comparison is not entirely one-sided, however. Those standalone purchases give you five complete albums, often with more elaborate jackets and, in several cases, clearly documented all-analog mastering. Coltrane 100 gives you selected performances from a much broader section of his career, and there are no previously unreleased recordings being promoted as part of the package. For anyone who already owns a substantial Coltrane collection, the question becomes less about price and more about duplication.

Who Is This For?

For someone beginning a serious Coltrane vinyl collection, the six-LP set makes a lot of sense because it provides almost four hours of music while connecting periods of his career that normally sit in different record bins and under different label logos. The chronological sequence also gives the listener context, which matters with Coltrane because the distance between his early sideman work and the final Impulse! recordings is enormous despite the relatively short span of time involved.

There is also a cheaper route. Rhino is releasing a two-LP Highlights From 1957-1964 edition for $34.98, featuring 11 selections including “Straight Street,” “Blue Train,” “Milestones,” “So What,” “Giant Steps,” “Naima,” “My Favorite Things,” “Equinox,” and “Acknowledgement.” A CD edition will sell for $14.98, which may be the smarter choice for listeners who want a well-selected introduction without committing to the full box.

Established collectors have a different calculation to make. If mastering provenance and owning complete albums are the priorities, I would still build the collection around the best OJC, Tone Poet, Acoustic Sounds Series, and Rhino High Fidelity releases. We have reviewed enough of those recently to know how good they can be, and the better individual editions offer a level of technical transparency that Rhino has not yet provided for this set.

The Bottom Line

Rhino’s John Coltrane 100 succeeds because it approaches Coltrane as a career rather than a collection of famous songs. Bringing Prestige, Blue Note, Atlantic, Columbia, and Impulse! material together creates a coherent path from the early sideman sessions through one of the most important artistic transformations in jazz history, and at $149.99 the six-LP set is priced fairly for what it includes.

Newer Coltrane listeners may find it more useful than buying five or six famous albums first and trying to connect the historical dots afterward. Longtime collectors should study the track list more carefully, because if Blue Train, Giant Steps, My Favorite Things, A Love Supreme, and the Prestige sessions are already sitting on your shelves in high-quality editions, much of the music will be familiar.

Where to buy: $149.99 at Amazon | Rhino Store

Related Reading:

Anthropic’s Claude AI is showing signs of overload again.

On Monday morning, users trying to work with Claude encountered failed requests and a “529 Overloaded” error. The message says the problem is a temporary server-side issue, then sends users to Anthropic’s status page. A screenshot shared with TNW shows the error appearing in the Claude interface.

Anthropic’s own incident report is broader, and more cautious. At 05:06 UTC on 24 August, the company said it was investigating elevated errors affecting requests to Claude Mythos 5, Claude Fable 5, Claude Opus 5, and Claude Opus 4.8. The incident also affects claude.ai, the Claude API, Claude Code, and Claude Cowork.

At 05:27 UTC, Anthropic said it had identified the cause and was working on a fix. Its latest update, posted at 06:42 UTC, says the company is still working to resolve the issue. The status page does not yet provide a recovery time or a technical explanation.

The distinction matters. “529 Overloaded” describes what at least some users are seeing, but Anthropic has not said that overload is the root cause of the wider incident. For now, the public record establishes elevated errors across several models and products, not why they occurred.

Claude’s busy August

This is not the first Claude service problem this month. Anthropic’s status history includes a degraded-performance incident affecting multiple models on 18 August, followed by another incident involving Opus 5 and Haiku 4.5 on 19 August. A separate disruption affecting multiple models was reported on 20 August.

Independent coverage has also documented the latest interruption. Frandroid reported that some users could not access Anthropic’s models and that the company had identified the problem. Online discussions on Hacker News and other forums show users posting the same 529 error while trying to use Claude.

Those reports do not establish the scale of the outage. They do show how quickly an infrastructure issue becomes visible: in a broken chat request, a failed coding session, or an API call that needs to be retried.

The cost of shared infrastructure

Claude is no longer one product. It is a family of models exposed through consumer chat, developer tools, an API, and Cowork. When several surfaces fail together, users do not experience separate incidents. They experience one unavailable dependency.

That is the more important story here. Anthropic has identified the cause, but it has not yet explained whether the problem sits in a shared service, model routing, capacity, or another part of the stack. Any of those explanations would be speculation at this stage.

The fix is still being monitored. Until Anthropic publishes more detail, the question is not simply when Claude comes back. It is whether users can build on it without treating failure as part of the interface.

For developers, that distinction may matter later. During the outage, the practical question is simpler: whether to wait, retry, switch providers, or close the laptop and play with the cat.

SoftBank plans a record 1 trillion yen retail bond sale, about $6.3bn, the largest by any issuer in Japan, to fund investment commitments to OpenAI. Analysts say banks were reluctant to take the risk, leaving the deal dependent on retail investors.

SoftBank has found a lender of last resort for its AI bet, and it is Japanese households. The group plans a record ¥1 trillion retail bond sale, about $6.3bn and the largest by any issuer in Japan, to fund its investment commitments to OpenAI.

The terms are built to be tempting. The seven-year bonds are expected to price on 4 September with an indicative coupon of 4.3% to 4.9%, and SoftBank says it expects an A rating from Japan Credit Rating Agency.

International raters see it differently. S&P has SoftBank at BB+, one notch below investment grade, having revised its outlook to stable from negative in July.

The reason it is going to retail is not enthusiasm. “Banks are finding it difficult to take on the risk given weak deposit growth, the credit rating and the seven-year duration, leaving the deal more reliant on retail investors,” said Yuuki Fukumoto of NLI Research Institute, in a pattern that has already forced Oracle to find anchors outside the banking system.

It is also becoming routine. This is SoftBank’s third retail bond of the year, after raising ¥418bn in April and ¥260bn in June.

Even ¥1 trillion does not close the gap. Bloomberg Intelligence’s Sharon Chen says a shortfall above $20bn remains after this issue, making offshore issuance likely in the near term.

The commitments behind it are enormous. SoftBank has pledged more than $60bn to OpenAI and has separately sought a $10bn margin loan secured against its stake in the company.

Everyone is raising at once. Alphabet, Meta, Microsoft and Amazon have committed nearly $2.4tn to AI-related spending, and Alibaba raised about $10bn in Hong Kong on Monday for chips and data centres.

A European issuer could not do this the same way. Bond prospectuses across the EEA routinely carry a prohibition on sales to retail investors, because the PRIIPs rules require a key information document that most issuers decline to produce.

That sits awkwardly with what Brussels says it wants. The Savings and Investments Union, adopted in March 2025, exists to move some of the €10tn European households hold, roughly 70% of it sitting in bank deposits, into capital markets, at a moment when Masayoshi Son calls bubble talk absurd.

Japan has now shown what mobilised savings look like in practice. A retail saver funding a seven-year bet on a company that does not yet know how the technology will pay.

Harvey has launched Tenet, its first proprietary model for legal work, post-trained on Kimi K3, the open-weight model from Chinese startup Moonshot. The company is backed by OpenAI, whose models Tenet is designed to displace inside Harvey’s own product.

Harvey has stopped renting the thing its business depends on. The legal software company has launched Tenet, its first proprietary model, after years of routing customer work through models from OpenAI, Anthropic and Google.

The base it chose is the detail worth stopping on. Tenet is post-trained on Kimi K3, the open-weight model released in July by the Chinese startup Moonshot, in work Harvey says it did with Fireworks AI.

Consider who is on the cap table. OpenAI is an investor in Harvey alongside Sequoia and Andreessen Horowitz, so a company valued at $15.5bn has built its flagship on a Chinese base to reduce its dependence on its own backer.

The commercial logic is unsubtle. Every model call Harvey makes on a customer’s behalf is an invoice from a rival, and owning the engine turns a variable cost into a fixed one.

Quality is the other argument. Cofounder Gabe Pereyra, formerly of Google DeepMind, says Harvey already routes tasks to whichever model suits them, and Tenet adds an option shaped around legal work specifically.

The training material was manufactured. Harvey hired lawyers, on staff and through firms including Mercor and Snorkel, to invent mock disputes and case files and then grade how well models reasoned through them.

Europe’s answer to Harvey is Swedish. Legora is chasing a $10bn valuation and selling to many of the same firms, which makes the choice of base model a competitive question as much as a technical one.

For a European buyer the AI Act is the next question. A company that modifies someone else’s general purpose model becomes the provider of the modified version only when the change is significant, with the Commission’s guidelines using roughly a third of the original training compute as the indicative marker.

A post-train almost certainly falls well below that. Which means most of the obligations stay upstream, and a European law firm deploying Tenet ends up relying on a documentation chain that begins in Beijing.

The licence has a threshold of its own. Kimi K3 permits derivative models, but requires a separate agreement with Moonshot for model-as-a-service operators above $20mn of revenue in any 12 months, and Harvey runs at more than $350mn annualised, in a market already shaped by free Chinese models.

The open-weight debate has mostly been about capability and price. Harvey has quietly turned it into a question about who your counterparty is when the work is covered by privilege.

Motorola is preparing to become the first major smartphone maker beyond Google to sell phones built for GrapheneOS, the privacy-focused Android alternative best known for its tight security requirements. The catch is that these will not be cheap phones.

According to a Mastodon post by GrapheneOS, Motorola and the GrapheneOS project are targeting a 2027 launch for their first supported devices, with the initial lineup expected to consist of premium flagships. The phones are also expected to cost more than Google’s Pixel devices, which have long been the only mainstream smartphones officially supported by GrapheneOS.

GrapheneOS is finally moving beyond Pixel phones

The partnership, announced earlier this year, marks a significant expansion for GrapheneOS. The open-source operating system is built around privacy and security, adding protections such as hardened memory management, granular network permissions, enhanced sandboxing and other defenses beyond standard Android.

Until now, Pixel phones have been the project’s primary home because Google’s hardware met GrapheneOS’ strict security requirements. Motorola’s upcoming devices will change that, with the company working alongside the project to support the software.

The price, however, could be steep. Ars Technica reports that the GrapheneOS team expects the first Motorola phones to be premium flagship devices, potentially priced above Google’s Pixel lineup. With the Pixel 11 starting at $900 and the Pixel 11 Pro XL reaching $1,300, Motorola’s privacy-focused models could easily enter four-figure territory.

This Mastodon thread adds more context to Motorola’s plans. The initial devices are expected to receive seven years of proper software updates and will likely launch with Snapdragon hardware. GrapheneOS has indicated that Qualcomm’s newer flagship chips offer improved support for memory tagging, isolated components and other hardware-level security features needed for the project.

A more expensive phone could buy an unusual kind of freedom

The partnership matters because GrapheneOS users have effectively been tied to one hardware manufacturer. Anyone wanting the operating system’s security features had to buy a compatible Pixel, whether they liked Google’s phones or not.

Motorola gives those users another option, even if the first devices will not target the budget market. The company’s premium phones could also eventually give GrapheneOS support to form factors Google does not currently offer, including foldables. Ars Technica notes that Motorola’s recent flagship focus has increasingly centered on foldable devices, making them a plausible candidate for the new software.

More important than the hardware itself is the shift in what GrapheneOS represents. A project once almost entirely dependent on Pixel hardware is beginning to build relationships with major Android manufacturers. The thread also reports that Motorola is dedicating engineering resources to the porting and maintenance work, while GrapheneOS will also contribute directly to firmware and software support. The project says the phones should meet or exceed its requirements, including seven years of updates and protection for features such as boot modes.

Motorola has not yet revealed the phones themselves or confirmed exact pricing. GrapheneOS support is expected to arrive around launch in 2027, with installation likely to work through the project’s web-based installer. The bigger question is whether privacy-conscious buyers will pay a premium for choice. GrapheneOS may finally be escaping the Pixel, but freedom, it seems, could come with a flagship-sized price tag.

Lanterns episode 2 is now live on HBO Max — and it introduces viewers to the most famous Green Lantern villain of all, Thaal Sinestro.

Long-time Green Lantern fans will already know plenty about this recurring antagonist, his relationship to Hal Jordan, and why he’s deemed such a threat to the Green Lantern Corps and wider galaxy. However, unless you’ve read the source material and/or seen the disastrous 2011 Ryan Reynolds-starring Green Lantern movie, many of you won’t know much, if anything, about Sinestro.